{"schema_version":1,"generated_at":"2026-11-20T02:29:47.174778+00:00","article_count":299,"articles":[{"id":"journals:10.1039/d6sc06681j","kind":"journals","source":"Chemical Science","title":"Automated multimodal mining of structured data locked in the figures and tables of flow chemistry literature","url":"https://doi.org/10.1039/d6sc06681j","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1039%2Fd6sc06681j","date":"2026-11-19T17:24:42Z","timestamp":1795109082,"categories":[],"topic_ids":[],"keywords":[],"matched_keywords":["machine learning","molecular"],"matched_tags":["machine_learning","generic_chem"],"doi":"10.1039/d6sc06681j","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Wenyuan Zhao","Xianzhu Zhong","Xavier Jouffroy","Yosuke Ashikari","Hiromichi V. Miyagishi","Mohmmad S. Qenawy","Simeng Wang","Yirui Peng","Mikail Raffano Munsyi","Aiichiro Nagaki"],"journal":"Chemical Science","publisher":"RSC","impact_factor":7.5,"abstract":"Flow chemistry literature locks much of its quantitative process data inside figures and tables that show molecular structures as images, including the low-yield and failed reactions that machine learning most needs.","source_metadata":{"short_name":"Chem. Sci.","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"journals:10.1039/d6sc06542b","kind":"journals","source":"Chemical Science","title":"Generative models in industrial drug discovery: perspectives from REINVENT","url":"https://doi.org/10.1039/d6sc06542b","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1039%2Fd6sc06542b","date":"2026-11-19T15:02:58Z","timestamp":1795100578,"categories":[],"topic_ids":[],"keywords":["Generative models","REINVENT"],"matched_keywords":["Generative models","REINVENT","drug","chemical"],"matched_tags":["generative_model","reinforcement_learning","generic_chem"],"doi":"10.1039/d6sc06542b","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Jon Paul Janet"],"journal":"Chemical Science","publisher":"RSC","impact_factor":7.5,"abstract":"Generative models have become established tools for exploring chemical space in early-stage drug discovery and are among the most important practical applications of AI for science. Numerous studies have reported...","source_metadata":{"short_name":"Chem. Sci.","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"feeds:https://macinchem.org/2026/11/19/chemistry-careers/","kind":"feeds","source":"Macs in Chemistry (Macinchem Blog)","title":"Chemistry careers","url":"https://macinchem.org/2026/11/19/chemistry-careers/","detail_url":"/chemradar/article?u=https%3A%2F%2Fmacinchem.org%2F2026%2F11%2F19%2Fchemistry-careers%2F","date":"2026-11-19T12:07:18+00:00","timestamp":1795090038,"categories":["Blog"],"topic_ids":[],"keywords":[],"matched_keywords":[],"matched_tags":[],"doi":null,"external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":[],"journal":null,"publisher":null,"impact_factor":null,"abstract":"","source_metadata":{"source":"Macs in Chemistry (Macinchem Blog)","published_utc":"2026-11-19T12:07:18+00:00","seen_at":"2026-11-19T14:20:39.947520+00:00"}},{"id":"preprints:2611.25452v1","kind":"preprints","source":"arXiv","title":"FamilySpace: From Emergent Chemical Semantics to Controllable Crystal Generation","url":"https://arxiv.org/abs/2611.25452v1","detail_url":"/chemradar/article?u=https%3A%2F%2Farxiv.org%2Fabs%2F2611.25452v1","date":"2026-11-19T03:32:41Z","timestamp":1795059161,"categories":["Design de novo"],"topic_ids":["denovo"],"keywords":["generative models","reinforcement learning","variational autoencoder","VAE"],"matched_keywords":["generative models","reinforcement learning","variational autoencoder","VAE","Chemical"],"matched_tags":["generative_model","reinforcement_learning","vae","generic_chem"],"doi":null,"external_id":"2611.25452v1","pdf_url":"https://arxiv.org/pdf/2611.25452v1","code_url":"https://github.com/hhw11/FamilySpace","code_host":"GitHub","authors":["Haowei Hua","Irwin King","Wanyu Lin"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"Latent crystal generative models typically combine a variational autoencoder (VAE) for learning atom-wise representations with a flow model that transports a simple source distribution to the learned latent distribution. However, previous studies have mainly focused on the quality of generated crystals, leaving the structure and generative utility of the atom-wise latent space largely unexplored. In this work, we show that even a vanilla crystal VAE implicitly organizes atom-wise latent representations according to elemental families, with different families occupying largely separated regions of the latent space. We term this phenomenon emergent chemical semantics. This structure suggests that sampling from a family-specific latent region can bias the decoder toward atoms from the corresponding elemental family. Building on this insight, we propose FamilySpace, a framework for controlling the elemental-family composition of generated crystals. FamilySpace assigns each elemental family a distinct source component, forming a family-indexed multi-source distribution, and trains a flow model using source-target pairs matched by family. In inference, a user-specified family template selects the source component for each atom-wise representation, enabling direct control over the family composition of generated crystals. Experiments on MP-20 and MPTS-52 show that FamilySpace achieves competitive crystal-generation performance while matching prescribed elemental-family compositions. Moreover, selected family templates increase the fraction of generated crystals predicted to be stable or metastable, demonstrating that latent-space composition control can support targeted exploration of promising chemical regions without additional optimization techniques such as reinforcement learning. Code is available at https://github.com/hhw11/FamilySpace","source_metadata":{"categories":["cs.LG","cond-mat.mtrl-sci"],"code_url":"https://github.com/hhw11/FamilySpace","code_status":"found"}},{"id":"journals:10.1021/acsomega.6c07978","kind":"journals","source":"ACS Omega","title":"3,4-Bis(benzyloxy)benzaldehyde-Derived\nHydrazones:\nSynthesis, α-Amylase-Selective Inhibition, Antioxidant\nProfiling, and In Silico Evaluation","url":"https://doi.org/10.1021/acsomega.6c07978","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1021%2Facsomega.6c07978","date":"2026-11-19T00:00:00+00:00","timestamp":1795046400,"categories":["Docking & Screening"],"topic_ids":["docking"],"keywords":["Molecular docking","IC50"],"matched_keywords":["Molecular docking","compounds","compound","drug","Molecular","IC50"],"matched_tags":["docking","generic_chem","bioactivity"],"doi":"10.1021/acsomega.6c07978","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Hanifa Fatullayev","Ismail Celik","Leyla Paşayeva"],"journal":"ACS Omega","publisher":"ACS","impact_factor":null,"abstract":"Carbohydrate-hydrolyzing enzymes (α-amylase and α-glucosidase) are clinically validated targets for controlling postprandial hyperglycemia in type 2 diabetes mellitus (T2DM), and oxidative stress is a recognized contributor to disease progression. Herein, a series of hydrazone Schiff bases derived from 3,4-bis(benzyloxy)benzaldehyde were synthesized by acid-catalyzed condensation of the aldehyde with para-substituted phenylhydrazines and were characterized by 1H and 13C NMR spectroscopy, FT-IR spectroscopy, and high-resolution mass spectrometry; chromatographic purity was established by RP-HPLC (94.5–98.6 area % at 350 nm). The series was evaluated for in vitro inhibition of porcine pancreatic α-amylase and yeast α-glucosidase and for antioxidant capacity using four complementary assays (DPPH, ABTS, FRAP, and CUPRAC). The compounds displayed a pronounced α-amylase preference: compound 9 (IC50 = 160.40 ± 2.05 μg/mL) and compound 7 (IC50 = 166.01 ± 2.64 μg/mL) were the most potent and exceeded acarbose under the assay conditions (IC50 = 232.66 ± 4.93 μg/mL); on a molar basis, both are essentially equipotent with the reference drug (345.2 ± 4.4 and 378.6 ± 6.0 μM versus 360.4 ± 7.6 μM), and IC50 values are reported in both units throughout (Table 1). α-Glucosidase inhibition was moderate, with the para-chloro analogue 3 being the most active (IC50 = 83.13 ± 6.53 μg/mL; 187.7 ± 14.7 μM). The selectivity index (IC50 α-glucosidase/IC50 α-amylase) spans 0.22 for 3 to 2.38 for 9, so a single para-substituent determines which carbohydrase is preferred. Antioxidant behavior was assay-dependent: compounds 1 and 4 delivered the highest ferric-reducing power (FRAP = 1854.0 ± 1.5 and 1786.7 ± 5.2 μmol Fe2+/g), and compound 5 gave the strongest CUPRAC response (280.95 ± 3.43 mg GAE/g). Molecular docking against α-amylase (PDB 1B2Y) and α-glucosidase (PDB 3TOP), together with SwissADME, QikProp, and ADMETlab profiling, provides a structural rationale for the observed substituent effects. These results define a clear, substituent-driven selectivity relationship within a lipophilic dibenzyloxy hydrazone framework and establish the DBZO-hydrazone scaffold as a tractable system for further optimization.","source_metadata":{"short_name":"ACS Omega","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"journals:10.1021/acs.jcim.6c03192","kind":"journals","source":"Journal of Chemical Information and Modeling","title":"Comment on “Hybrid\nGraph–Machine Learning\nFramework for Accurate and Interpretable Band Gap Prediction”","url":"https://doi.org/10.1021/acs.jcim.6c03192","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1021%2Facs.jcim.6c03192","date":"2026-11-19T00:00:00+00:00","timestamp":1795046400,"categories":["Property Prediction"],"topic_ids":["property"],"keywords":["graph neural networks","GNN","DFT"],"matched_keywords":["neural networks","graph neural networks","GNN","Machine Learning","DFT"],"matched_tags":["neural_network","gnn","machine_learning","dft"],"doi":"10.1021/acs.jcim.6c03192","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Shambhu Bhandari Sharma"],"journal":"Journal of Chemical Information and Modeling","publisher":"ACS","impact_factor":5.3,"abstract":"Aydin et al. reported R2 = 0.921 for band gap prediction using hybrid graph neural networks. Reproduction shows this performance is inflated by data leakage from post-DFT proxies (is_metal, is_gap_direct) and embedding contamination. Controlled ablations reveal a genuine GNN gain of 0.08–0.11 in R2, while leakage contributes ∼0.06, lowering honest performance to 0.82–0.86 R2 (MAE 0.28–0.30 eV).","source_metadata":{"short_name":"JCIM","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"preprints:10.26434/chemrxiv.15002852/v5","kind":"preprints","source":"ChemRxiv","title":"De Novo Design of Enantioselective Artificial Metalloenzyme Photocatalysts","url":"https://doi.org/10.26434/chemrxiv.15002852/v5","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.26434%2Fchemrxiv.15002852%2Fv5","date":"2026-11-19T00:00:00Z","timestamp":1795046400,"categories":[],"topic_ids":[],"keywords":["De Novo Design"],"matched_keywords":["De Novo Design"],"matched_tags":["de_novo_design"],"doi":"10.26434/chemrxiv.15002852/v5","external_id":"10.26434/chemrxiv.15002852/v5","pdf_url":null,"code_url":null,"code_host":null,"authors":["Sandip Mishra","Payal","Declan Evans","Kingsley Bortey","Husayn Bootwala","Giovanni Gonzalez-Gutierrez","Ricardo Javier Vázquez","David Baker","Jared C Lewis"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"","source_metadata":{"date_field":"posted","keywords":[]}},{"id":"journals:10.1021/acs.jcim.6c02673","kind":"journals","source":"Journal of Chemical Information and Modeling","title":"DeepChem-DEL: An\nOpen-Source Framework for Reproducible\nDEL Modeling and Benchmarking","url":"https://doi.org/10.1021/acs.jcim.6c02673","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1021%2Facs.jcim.6c02673","date":"2026-11-19T00:00:00+00:00","timestamp":1795046400,"categories":["Cheminformatics","Docking & Screening","Design de novo","Library Design"],"topic_ids":["cheminformatics","docking","denovo","libraries"],"keywords":["ChemBERTa","random forest","denoising","DNA encoded libraries"],"matched_keywords":["ChemBERTa","random forest","Machine learning","ML","denoising","DNA-encoded libraries","chemical","compounds"],"matched_tags":["pretrained_chem_models","ensemble","machine_learning","diffusion","dna_encoded_library","generic_chem"],"doi":"10.1021/acs.jcim.6c02673","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Aryan Amit Barsainyan","Riya Singh","Abhiraj Pravin Mengade","Rida Irfan","Bharath Ramsundar"],"journal":"Journal of Chemical Information and Modeling","publisher":"ACS","impact_factor":5.3,"abstract":"DNA-encoded libraries (DELs) allow screening of extremely large chemical spaces by tagging compounds with DNA barcodes. Machine learning (ML) techniques have enabled the extraction of meaningful chemical patterns from noisy, target-specific DEL binding data. However, existing tools for DEL machine learning (DEL-ML) remain fragmented, which limits systematic construction and reproducibility of modeling pipelines. Moreover, this fragmentation hinders integration of new models, such as chemical foundation models, into DEL workflows alongside classical approaches. To address these challenges in DEL modeling, we introduce DeepChem-DEL, an open-source suite of reproducible ML workflows for DEL modeling built on the DeepChem ecosystem. DeepChem-DEL integrates (i) a configurable denoising pipeline with (ii) modular DeepChem workflows for enrichment prediction, hit identification, and benchmarking across diverse model classes, including chemical foundation models enabled through extensions to the open-source ChemBERTa-3 framework. We evaluate DeepChem-DEL on the KinDEL and BELKA data sets, reproducing key baselines and enabling direct comparison across model classes. Our experiments demonstrate that leveraging chemical foundation models yields performance competitive with, and in several cases exceeding, previously explored approaches such as graph convolutional networks and random forest models. Together, these results demonstrate that DeepChem-DEL is a scalable and extensible platform for DEL-ML, while highlighting the potential of chemical foundation models for improved DEL modeling and prioritization.","source_metadata":{"short_name":"JCIM","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"preprints:10.26434/chemrxiv.15011551/v2","kind":"preprints","source":"ChemRxiv","title":"Design and Benchmarking of CK2α′-Preferring Inhibitors for Huntington’s Disease: The Limits of Structure-Based Isoform Selectivity","url":"https://doi.org/10.26434/chemrxiv.15011551/v2","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.26434%2Fchemrxiv.15011551%2Fv2","date":"2026-11-19T00:00:00Z","timestamp":1795046400,"categories":[],"topic_ids":[],"keywords":["ChEMBL","kinase"],"matched_keywords":["ChEMBL","compounds","Molecular","Protein","kinase"],"matched_tags":["chem_benchmarks","generic_chem","biology"],"doi":"10.26434/chemrxiv.15011551/v2","external_id":"10.26434/chemrxiv.15011551/v2","pdf_url":null,"code_url":null,"code_host":null,"authors":["Tarun Tarun"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"Protein kinase CK2 acts through two catalytic isoforms, CK2α (A1) and CK2α′ (A2), whose ATP-binding sites differ at only two hinge positions, making isoform-selective inhibition difficult. We tested whether a standard structure-based pipeline can design and rank CK2α′-preferring compounds. In a paired benchmark of 210 compounds measured against both isoforms, the inter-isoform docking score was unrelated to the experimental selectivity (Spearman ρ = −0.22), and even within the benzo[c][2,6]naphthyridine scaffold, the scoring reversed the order set by the ChEMBL activities. The only reliable determinant of preference, a two-carbon linker between the anilino nitrogen and the pendant ring of this scaffold, emerged from the measured activities rather than from any computed score. Molecular-dynamics ensembles of the two isoforms overlapped (0.88 Å between versus 0.72–0.96 Å within), placing the selectivity signal below the resolution of the methods. From an enumerated series, EV043 showed the largest end-point free-energy preference for CK2α′ (ΔΔG b -7.74 kcal/mol), although this value remained sensitive to the starting pose in CK2α, while the non-selective inhibitor CX-4945 was reproduced correctly (+0.08 kcal/mol). Distinguishing CK2α′ from CK2α therefore requires alchemical free-energy methods supported by experimental validation.","source_metadata":{"date_field":"posted","keywords":[]}},{"id":"journals:10.1021/acs.jafc.6c11247","kind":"journals","source":"Journal of Agricultural and Food Chemistry","title":"Design, Synthesis,\nand Antifungal Evaluation of Novel\nPyridine-3-amide Derivatives Containing Phenylthiazole Moiety as Potential\nSuccinate Dehydrogenase Inhibitors","url":"https://doi.org/10.1021/acs.jafc.6c11247","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1021%2Facs.jafc.6c11247","date":"2026-11-19T00:00:00+00:00","timestamp":1795046400,"categories":["Docking & Screening"],"topic_ids":["docking"],"keywords":["Molecular docking","EC50"],"matched_keywords":["Molecular docking","compound","Molecular","EC50"],"matched_tags":["docking","generic_chem","bioactivity"],"doi":"10.1021/acs.jafc.6c11247","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Jin Wang","Zhiyuan Xu","Sen Huang","Huochun Ye","Fadi Zhu","Jing Zhang","Gang Feng"],"journal":"Journal of Agricultural and Food Chemistry","publisher":"ACS","impact_factor":null,"abstract":"In search of novel succinate dehydrogenase inhibitor fungicides (SDHIs) for the management of plant fungal diseases, 29 novel pyridine-3-amide derivatives containing a phenylthiazole moiety were designed and synthesized. Among them, compound H18 showed the strongest antifungal activity against B. cinerea, with an EC50 value of 0.06 μg/mL, substantially lower than that of boscalid (1.79 μg/mL). Spearman’s correlation analysis revealed no cross-resistance between H18 and boscalid (ρ = −0.1743, P = 0.6082). Moreover, H18 exhibited strong protective and curative efficacy of 79.98% and 75.57%, respectively, against B. cinerea on tomatoes at a concentration of 100 μg/mL, superior to that of boscalid. Mechanistic studies demonstrated that H18 inhibited SDH activity, leading to ROS accumulation, disruption of mitochondrial energy metabolism, and fungal cell death. Molecular docking suggested that the flexible conformations of the phenylthiazole moiety in H18 may contribute to its potential resistance-overcoming capacity. These findings identify H18 as a promising new SDHI candidate for B. cinerea control, warranting further investigation.","source_metadata":{"short_name":"J. Agric. Food Chem.","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"journals:10.1021/acs.chemrev.6c00374","kind":"journals","source":"Chemical Reviews","title":"Development\nand Application of Nanopore Direct RNA\nSequencing","url":"https://doi.org/10.1021/acs.chemrev.6c00374","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1021%2Facs.chemrev.6c00374","date":"2026-11-19T00:00:00+00:00","timestamp":1795046400,"categories":[],"topic_ids":[],"keywords":[],"matched_keywords":["machine learning","protein"],"matched_tags":["machine_learning","biology"],"doi":"10.1021/acs.chemrev.6c00374","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Miten Jain","Robin Abu-Shumays","Hugh E. Olsen","Mark Akeson"],"journal":"Chemical Reviews","publisher":"ACS","impact_factor":55.8,"abstract":"Individual RNA strands can be sequenced directly using nanoscale protein pores. This process, Nanopore Direct RNA Sequencing (Nanopore DRS), uses a helicase motor protein to regulate RNA displacement through the nanopore orifice in single nucleotide steps. Partial blockade of the monovalent ionic current is recorded at 4 kHz and converted to a string of RNA nucleotide basecalls using machine learning models. In this review, we discuss the development of Nanopore DRS from 1996 to the present. This includes performance metrics, evolution of machine learning models, and applications of Nanopore DRS to major RNA classes across kingdoms. Advantages of the technology are highlighted with special emphasis on nucleotide modifications along intact single strands. Shortcomings of the technology are also considered along with possible solutions. We close with forward looking Nanopore DRS applications: Single cell RNA sequencing; remote monitoring of infectious agents; principled understanding of nanopore ionic current patterns using all-atoms simulations; global food security; and clinical applications.","source_metadata":{"short_name":"Chem. Rev.","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"journals:10.1038/s41598-026-81415-y","kind":"journals","source":"Scientific Reports","title":"Evaluation of sensing response of H2S sensing materials toward H2S molecules using artificial neural network and response surface methodology","url":"https://doi.org/10.1038/s41598-026-81415-y","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1038%2Fs41598-026-81415-y","date":"2026-11-19T00:00:00+00:00","timestamp":1795046400,"categories":[],"topic_ids":[],"keywords":[],"matched_keywords":["neural network","molecules"],"matched_tags":["neural_network","generic_chem"],"doi":"10.1038/s41598-026-81415-y","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Fatemeh Hamidipour","Hadiseh Masoumi","Ahmad Rahbar-Kelishami"],"journal":"Scientific Reports","publisher":"Nature","impact_factor":null,"abstract":"","source_metadata":{"short_name":"Sci. Rep.","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"preprints:10.26434/chemrxiv.15012462/v1","kind":"preprints","source":"ChemRxiv","title":"Exploring Boltzmann Distributions via Kernel-Based Fokker–Planck Ensemble Transport","url":"https://doi.org/10.26434/chemrxiv.15012462/v1","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.26434%2Fchemrxiv.15012462%2Fv1","date":"2026-11-19T00:00:00Z","timestamp":1795046400,"categories":[],"topic_ids":[],"keywords":["Transformer","free energy"],"matched_keywords":["Transformer","free energy"],"matched_tags":["transformer","thermodynamics"],"doi":"10.26434/chemrxiv.15012462/v1","external_id":"10.26434/chemrxiv.15012462/v1","pdf_url":null,"code_url":null,"code_host":null,"authors":["Xiaoxuan Yu","Zichen Sun","Jun Zhang","Yi Qin Gao"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"Local trajectory-based samplers can relax within a visited basin long before finite trajectories cross barriers to other Boltzmann-relevant regions. We introduce the Fokker–Planck Transformer (FPT), a training-free interacting-ensemble method that makes the instantaneous ensemble distribution a dynamical variable, shifting exploration from trajectory-wise barrier crossing toward distribution-level redistribution. FPT discretizes the current-velocity representation of overdamped Fokker–Planck evolution by combining energetic drift with a kernel estimate of the ensemble score. The continuum flow has the Boltzmann density as its stationary target and monotonically dissipates configurational free energy and relative entropy. A Gaussian row-mass selector couples increasing local occupancy to a vanishing spatial scale without a separate bandwidth limit. For its fresh explicit-Euler realization, pathwise score residuals propagate to distributional and observable errors, while a weak-flux argument gives conditional fixed-horizon convergence to the Fokker–Planck solution. For this realization, a controlled two-dimensional calculation tracks transport from a localized Gaussian toward equilibrium against a numerically propagated Fokker–Planck reference. In a three-dimensional periodic landscape, an exact Fokker–Planck reference shows that a shared-bandwidth approximation populates distant wells far ahead of the continuum solution, which accounts for its faster support discovery; the transported ensemble runs warm, and a Hastings-corrected continuation from its endpoint restores the energy distribution within two steps. Further tests in a 100-dimensional landscape, alanine decapeptide, and 190 docking systems show rapid expansion of Boltzmann-relevant support, with the clearest gains over matched local controls when barriers or configurational complexity limit global exploration. Overall, FPT provides thermodynamically grounded finite-budget support discovery and can seed target-preserving sampling.","source_metadata":{"date_field":"posted","keywords":[]}},{"id":"journals:10.1038/s43588-026-01082-x","kind":"journals","source":"Nature Computational Science","title":"Generative molecular design needs oracle grounding","url":"https://doi.org/10.1038/s43588-026-01082-x","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1038%2Fs43588-026-01082-x","date":"2026-11-19T00:00:00+00:00","timestamp":1795046400,"categories":["Design de novo"],"topic_ids":["denovo"],"keywords":["Generative molecular design"],"matched_keywords":["Generative molecular design","molecular"],"matched_tags":["generative_drug_design","generic_chem"],"doi":"10.1038/s43588-026-01082-x","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Sarina Kopf","Philippe Schwaller"],"journal":"Nature Computational Science","publisher":"Nature","impact_factor":18.3,"abstract":"","source_metadata":{"short_name":"Nat. Comput. Sci.","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"journals:10.1021/acsomega.6c10360","kind":"journals","source":"ACS Omega","title":"In Silico\n                    Identification of Peptide\nPharmacophoric Determinants and Complementary Binding Hotspots on\nFungal Grr1","url":"https://doi.org/10.1021/acsomega.6c10360","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1021%2Facsomega.6c10360","date":"2026-11-19T00:00:00+00:00","timestamp":1795046400,"categories":["Docking & Screening","Free Energy & MD"],"topic_ids":["docking","freeenergy"],"keywords":["pharmacophore modeling","pharmacophore mapping","binding free energy","molecular features","molecular dynamics"],"matched_keywords":["pharmacophore modeling","pharmacophore mapping","binding free energy","molecular features","molecular dynamics","molecular","ligand","ligands","protein"],"matched_tags":["pharmacophore","thermodynamics","molecular_descriptor","molecular_dynamics","generic_chem","biology"],"doi":"10.1021/acsomega.6c10360","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Mateus Dias-Oliveira","Henrique Bridi","Rogerio Margis","Geancarlo Zanatta"],"journal":"ACS Omega","publisher":"ACS","impact_factor":null,"abstract":"The growing burden of human and plant fungal diseases, compounded by limited therapeutic options and emerging antifungal resistance, underscores the need to identify novel molecular targets for antifungal intervention. In this context, the fungal F-box protein Grr1 acts as the substrate-recognition component of the SCF ubiquitin ligase complex, binding phosphorylated degrons to direct substrate degradation and regulate key cellular processes. However, the structural and energetic determinants governing phosphodegron recognition by the Grr1 leucine-rich repeat (LRR) domain remain poorly characterized. Here, we investigated the molecular basis of Grr1 recognition of phosphodegrons derived from the substrates Cln2 and Gic2 and evaluated whether phosphomimetic substitutions could preserve their interaction with the LRR domain. Structural modeling, molecular dynamics simulations, binding free-energy calculations, per-residue energy decomposition, and pharmacophore modeling were integrated to characterize the Grr1-degron interfaces and identify their major recognition determinants. Phosphorylation was required to maintain stable association of both degrons with the concave surface of the Grr1 LRR domain. Per-residue energy analysis revealed a conserved basic recognition surface comprising Arg418, Arg496, Lys498, Arg550, Lys608, Arg680, and Arg709, with Lys498, Arg550, Arg680, and Arg709 emerging as the principal energetic contributors shared by both complexes. Together, these residues form an electropositive environment that coordinates the degron phosphate groups, while pharmacophore mapping identified complementary molecular features associated with ligand recognition. Remarkably, substitution of the phosphorylated residues with aspartate preserved the calculated binding free energy of both degrons, whereas glutamate substitution disrupted stable binding, indicating that phosphomimetic recognition depends not merely on negative charge but critically on its precise spatial positioning within the recognition interface. Collectively, these findings define a conserved structural and energetic mechanism of phosphodegron recognition by fungal Grr1 and identify molecular features that can be exploited for ligand development. This framework supports the rational design of phosphomimetic peptides and other competitive ligands targeting the Grr1 substrate-recognition interface and provides a foundation for exploring this conserved regulatory system as a potential strategy for antifungal intervention.","source_metadata":{"short_name":"ACS Omega","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"journals:10.1038/s41598-026-78564-5","kind":"journals","source":"Scientific Reports","title":"In silico screening of FDA-approved drugs for elimination of senescent cells via senescence-associated anti-apoptotic pathway inhibition","url":"https://doi.org/10.1038/s41598-026-78564-5","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1038%2Fs41598-026-78564-5","date":"2026-11-19T00:00:00+00:00","timestamp":1795046400,"categories":[],"topic_ids":[],"keywords":["In silico screening"],"matched_keywords":["In silico screening","drugs"],"matched_tags":["virtual_screening","generic_chem"],"doi":"10.1038/s41598-026-78564-5","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Emad Manni"],"journal":"Scientific Reports","publisher":"Nature","impact_factor":null,"abstract":"","source_metadata":{"short_name":"Sci. Rep.","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"journals:10.1021/acs.jafc.6c10849","kind":"journals","source":"Journal of Agricultural and Food Chemistry","title":"Integrative Omics\nReveal the Antiviral Mechanism of\nGossypol and Guide the Construction of a Chitooligosaccharide–Gossypol\nNanocomposite","url":"https://doi.org/10.1021/acs.jafc.6c10849","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1021%2Facs.jafc.6c10849","date":"2026-11-19T00:00:00+00:00","timestamp":1795046400,"categories":["Docking & Screening"],"topic_ids":["docking"],"keywords":["Molecular docking","molecular dynamics"],"matched_keywords":["Molecular docking","Molecular","molecular dynamics"],"matched_tags":["docking","generic_chem","molecular_dynamics"],"doi":"10.1021/acs.jafc.6c10849","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Wenkang Li","Weifeng Wu","Huimin Chang","Shuo Qin","Maozhi Ren","Peng Wang","Yan Zhao","Jun Li","Fuguang Li","Wenjing Wang"],"journal":"Journal of Agricultural and Food Chemistry","publisher":"ACS","impact_factor":null,"abstract":"Gossypol (GOS), a unique cotton sesquiterpenoid, exhibits broad antiviral activity against animal viruses, but its specific effects on plant viruses remain largely unexplored. Using tobacco mosaic virus (TMV) and cucumber mosaic virus (CMV) models, we demonstrated that GOS treatment significantly reduces viral RNA accumulation and alleviates systemic symptoms. Integrating multiple omics approaches, findings suggest that GOS rebalances host hormones by repressing biosynthesis and activating catabolism, curbing excess salicylic acid (SA) and abscisic acid (ABA) while restoring indole acetic acid (IAA) to optimize resource allocation between growth and defense. Additionally, the transcriptional upregulation of antiviral sesquiterpene genes closely aligns with the targeted metabolite accumulation. Molecular docking, molecular dynamics, and surface plasmon resonance (SPR) assays reveal that GOS directly targets viral RNA-dependent RNA polymerase (RdRp). Collectively, GOS mitigates plant viral infections through polymerase targeting and defense reprogramming. These insights inspired a highly efficient GOS–chitooligosaccharide (COS) nanocomplex for crop protection in sustainable agriculture.","source_metadata":{"short_name":"J. Agric. Food Chem.","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"preprints:10.26434/chemrxiv.15012474/v1","kind":"preprints","source":"ChemRxiv","title":"Linear-Algebraic Audit of Biomedical Matrices: Five Empirical Case Studies in Regularized Inference","url":"https://doi.org/10.26434/chemrxiv.15012474/v1","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.26434%2Fchemrxiv.15012474%2Fv1","date":"2026-11-19T00:00:00Z","timestamp":1795046400,"categories":[],"topic_ids":[],"keywords":["chemometric"],"matched_keywords":["chemometric"],"matched_tags":["chemometrics"],"doi":"10.26434/chemrxiv.15012474/v1","external_id":"10.26434/chemrxiv.15012474/v1","pdf_url":null,"code_url":null,"code_host":null,"authors":["Petr Avtonomov"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"High-throughput biomedical assays—ranging from gene expression microarrays to mass spectrometry and clinical registries—routinely exhibit severe rank deficiency (N ≪ P ), zero-inflation, and extreme sample outliers. Standard chemometric and bioinformatics pipelines apply variance auto-scaling, simplex normalization, and bilinear projections (PCA/PLS-DA) without verifying operator conditioning or modal localization. Grounded in inverse problem theory, this empirical study subjects five public biomedical datasets to a sequential four-step linear-algebraic audit evaluating residual Gram conditioning (κ), Effective Degrees of Freedom (DoF), singular spectra (Σ), and Participation Ratios (PR). Across five empirical regimes—Golub leukemia microarrays (72×7129), a large-scale oncology cohort (1545×10935), zero-inflated NCI Arcene mass spectra (200 × 10000), Pomeroy central nervous system tumor profiles (60 × 7129), and the Cleveland cardiology registry (270 × 13)—we demonstrate how uncalibrated preprocessing inflates background noise threefold, collapses leading components onto fewer than 7 outlier samples, and bends multi-batch artifacts into parabolic horseshoes. Removing just 0.6 percent of leverage outliers identified by PR restores sample coherence nearly ninety-fold (PR : 6.4 → 548.1), providing a rigorous diagnostic standard for biomedical data hygiene.","source_metadata":{"date_field":"posted","keywords":[]}},{"id":"journals:10.1038/s41598-026-80158-0","kind":"journals","source":"Scientific Reports","title":"Machine learning and molecular validation reveal the PI3K/AKT/mTOR inflammatory axis for therapeutic targeting in severe COVID-19","url":"https://doi.org/10.1038/s41598-026-80158-0","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1038%2Fs41598-026-80158-0","date":"2026-11-19T00:00:00+00:00","timestamp":1795046400,"categories":[],"topic_ids":[],"keywords":[],"matched_keywords":["Machine learning","molecular"],"matched_tags":["machine_learning","generic_chem"],"doi":"10.1038/s41598-026-80158-0","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Sahar Samsami","Hiva Alipanah","Alireza Faghani Ghodrat","Mohammad Mehdi Naghizadeh","Seyedeh-Samira Ashrafmansouri","Ali Ghanbariasad"],"journal":"Scientific Reports","publisher":"Nature","impact_factor":null,"abstract":"","source_metadata":{"short_name":"Sci. Rep.","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"journals:10.1038/s41598-026-77567-6","kind":"journals","source":"Scientific Reports","title":"Rapid quantification of patchouli alcohol in patchouli oil (Pogostemon Cablin) using near infrared spectroscopy with multivariate and machine learning approaches","url":"https://doi.org/10.1038/s41598-026-77567-6","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1038%2Fs41598-026-77567-6","date":"2026-11-19T00:00:00+00:00","timestamp":1795046400,"categories":["Spectra & Analytical"],"topic_ids":["chemometrics"],"keywords":[],"matched_keywords":["infrared spectroscopy with multivariate and machine learning","machine learning"],"matched_tags":["analytical_chemistry_ai","machine_learning"],"doi":"10.1038/s41598-026-77567-6","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Nurul Ainina","Sarifah Nurjanah","Uma Rani Sinniah","Rahmat Hidayat","Euis Julaeha","Muhammad Achirul Nanda","Mimin Muhaemin"],"journal":"Scientific Reports","publisher":"Nature","impact_factor":null,"abstract":"Patchouli oil ( Pogostemon Cablin ) is an essential oil primarily used as a fixative ingredient in the perfumery and cosmetic industries. Its major component, patchouli alcohol (PA), is recognized as a key quality parameter, with higher contents being desirable. However, the conventional method for quantifying PA, namely gas chromatography-mass spectrometry (GC-MS), is relatively expensive, time-consuming, and destructive. Therefore, this study aimed to explore near-infrared (NIR) spectroscopy combined with multivariate and machine learning algorithms, including partial least squares regression (PLSR), support vector regression (SVR), and a hybrid method, partial least squares-support vector regression (PLS-SVR). The best-performing models was obtained using PLSR combined with Savitzky-Golay 2nd derivative (SG 2nd) preprocessing in a single-split evaluation, achieving R 2 CV = 0.9094, RMSECV = 2.3471, R 2 P = 0.9251, RMSEP = 2.3099, and RPD = 3.676. The robustness of this model was further confirmed using Monte-Carlo double cross-validation (MCCV-DCV), yielding a mean of R 2 P and RMSEP at 0.8909 ± 0.0417 and 2.4329 ± 0.4402, respectively, across thirty repeated iterations. A paired statistical comparison between PLSR and PLS-SVR further showed no significant difference in RMSEP (Wilcoxon singed-rank test, p-value = 0.1347), while SVR performed significantly worse ( p < 0.05). In conclusion, NIR spectroscopy using PLSR method, could be used as a rapid, simple, low-cost alternative method for estimating PA content in patchouli oil.","source_metadata":{"short_name":"Sci. Rep.","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"preprints:10.26434/chemrxiv.15012472/v1","kind":"preprints","source":"ChemRxiv","title":"Self-Consistent Induced Dipoles for MACE-POLAR in Molecular Embedding and Classical Dynamics","url":"https://doi.org/10.26434/chemrxiv.15012472/v1","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.26434%2Fchemrxiv.15012472%2Fv1","date":"2026-11-19T00:00:00Z","timestamp":1795046400,"categories":["ML Potentials"],"topic_ids":["potentials"],"keywords":["MACE","DFT"],"matched_keywords":["MACE","DFT","Molecular"],"matched_tags":["mlip_tools","dft","generic_chem"],"doi":"10.26434/chemrxiv.15012472/v1","external_id":"10.26434/chemrxiv.15012472/v1","pdf_url":null,"code_url":null,"code_host":null,"authors":["Jianxin Cheng"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"We couple the native Gaussian multipoles of MACE-POLAR-1-M to classical induced dipoles and compare the resulting solute densities, fields and forces with quantum references. The density-driven fixed point has a nonzero energy derivative with respect to its dipoles. An adjoint recovers forces on the converged surface to within 2.13 × 10−7 eV/Å of reconverged finite differences in twelve force-test configurations. A separate density comparison treats both partners quantum mechanically at twelve geometries of six small complexes. At a common Gaussian width of 1.5 Å, enabling mutual induction changes the native MACE solute density by at most 6.81 × 10−4 𝑒 in integrated absolute difference. For water facing chloride at 3.15 Å, the density change is 0.000107 𝑒 with the MACE-induced dipole, 0.00426 𝑒 with the prescribed DFT-induced dipole, and 0.01026 𝑒 with DFT at that same dipole. Thus the error involves both the field that induces chloride and the water response to the returning field. For Na+–Ar at 6 Å, the predicted solute multipoles are unchanged, while environment induction gives an interaction of −9.19 meV, close to the classical induction limit","source_metadata":{"date_field":"posted","keywords":[]}},{"id":"journals:10.1021/acs.jcim.6c02968","kind":"journals","source":"Journal of Chemical Information and Modeling","title":"Shake It Off: Abl\nKinase Regulatory Domain Dynamics\nExplain Conformational Changes upon Activation","url":"https://doi.org/10.1021/acs.jcim.6c02968","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1021%2Facs.jcim.6c02968","date":"2026-11-19T00:00:00+00:00","timestamp":1795046400,"categories":["Free Energy & MD"],"topic_ids":["freeenergy"],"keywords":[],"matched_keywords":[],"matched_tags":["molecular_dynamics","biology","generic_chem"],"doi":"10.1021/acs.jcim.6c02968","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Philipe Oliveira Fernandes","José Luis Muñoz Reyes","Pedro Tenório T. F. Leite","Diego Magno Martins","Antti Poso","Ina Pöhner","Adolfo Henrique Moraes","Vinicius G. Maltarollo","Thales Kronenberger"],"journal":"Journal of Chemical Information and Modeling","publisher":"ACS","impact_factor":5.3,"abstract":"Abl kinase is a major molecular target for chronic myeloid leukemia therapy. Its activity is tightly controlled by its regulatory module, composed of the SH3 and SH2 domains and the SH2–kinase linker. Although the assembled autoinhibited and extended active conformations have been structurally characterized, the molecular pathway connecting these states remains poorly understood. Simulations initiated from the end point structures failed to sample the complete transition, even when Gaussian accelerated molecular dynamics (GaMD) was employed to enhance conformational sampling. To overcome this limitation, we combined multiscale modeling with extensive molecular dynamics simulations and Markov state modeling to reconstruct the Abl disassembly pathway. The resulting kinetic landscape revealed two coupled conformational equilibria: a faster exchange between the canonical assembled state and an intermediate ensemble that captures the initial rotation of the regulatory module, and a slower transition toward the extended state, characterized by migration of the regulatory module atop the kinase domain N-lobe. The assembled ensemble is favored (∼88% of the equilibrium population), with the canonical autoinhibited arrangement alone accounting for approximately 60%, while early disassembly disrupts the E98–K294 and SH3–linker interactions and releases Y226 from its protective interactions, while Y70 remains accessible through the intrinsic dynamics of the SH3 domain. By revealing how these phosphorylation-sensitive residues become accessible during the earliest stages of disassembly, our model provides a mechanistic framework for Abl activation and identifies regulatory interfaces as potential targets for allosteric stabilization of the autoinhibited state.","source_metadata":{"short_name":"JCIM","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"preprints:10.26434/chemrxiv.15012473/v1","kind":"preprints","source":"ChemRxiv","title":"Towards a Metrological Validation of PFAS Property Predictions by In-Context LLMs Across Four Architectures","url":"https://doi.org/10.26434/chemrxiv.15012473/v1","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.26434%2Fchemrxiv.15012473%2Fv1","date":"2026-11-19T00:00:00Z","timestamp":1795046400,"categories":["Property Prediction","LLMs & Agents"],"topic_ids":["property","llm"],"keywords":["QSPR","LLMs","LLM","DFT","dissociation constant","partition coefficient"],"matched_keywords":["QSPR","LLMs","LLM","DFT","dissociation constant","partition coefficient","chemical","molecular","molecules"],"matched_tags":["qsar_qspr","LLMs","dft","inhibition_constant","solubility","generic_chem"],"doi":"10.26434/chemrxiv.15012473/v1","external_id":"10.26434/chemrxiv.15012473/v1","pdf_url":null,"code_url":null,"code_host":null,"authors":["Carlos Gonzalez"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"In recent years, the use of large-language models (LLMs) as prediction tools has become increasingly popular. Despite this popularity, the metrological validation of LLMs in the study of chemical problems has been somehow neglected to date. To illustrate this need in chemistry, we introduce a minimal, metrologically grounded framework for validating in-context learning (ICL)-based prediction in LLMs on molecular physicochemical properties with reliable estimation of uncertainty. Our framework integrates a multicycle testing protocol (Mc = 50 replicate cycles per model-property cell) with a six-metric bundle: central tendency (mean MAE over replicate cycles), reproducibility (interquartile range), regression performance (RMSE, and R2), information-theoretical lower bound on the amount of captured signal (𝐼\" in nats), and ordinal 2 performance (accuracy against the median in the dataset), plus three cross-property and crossmodel comparison metrics: NRMSE, Welch’s z-test, and Scaling Law Fit exponents. We use this framework to validate four LLMs, two dense (8B and 70B parameters; 1B = 109) and two mixture-of-experts (MoE; 17B active parameters with 109B and 400B in total), for predicting eight per- and polyfluoroalkyl substance (PFAS) properties: boiling point (BP), melting point (MP), vapor pressure (VP), Henry’s law constant (HLC), octanol-water partition coefficient (LogKow), aqueous solubility (LogS), acid dissociation constant (pKa), and critical micelle concentration (LogCMC). In this benchmark, the frontier MoE model (Maverick 400B parameters) achieves a mean validation NRMSE of 14.4 %, and the mid-scale MoE model (Scout, 109B parameters) outperforms our larger dense model (Llama-3.3-70B) despite having 4.1× fewer active parameters. This suggests that MoE routed capacity can substitute for activeparameter scale in chemical applications of ICL. When compared to published QSPR and DFT baselines, LLM-ICL beats these approaches on three of eight directly comparable properties (pKa by 43 %, LogKow by 31 %, HLC by 10 %) and trails behind on five (BP, MP, LogS, VP by 12 % to 75 %, and LogCMC by a factor of approx. 3.3), despite using approximately 32 incontext examples (32 for seven properties, 20 for HLC) versus the training sets comprising 102 to 105 molecules used in these approaches. We also discuss a validation-set anomaly in the case of LogKow, where our mid-scale dense model Llama-3.3-70B outperforms the frontier MoE model on all four validation metrics, possibly due to better coverage of the pretraining data. Permutation (shuffled-label) and information-access blinding controls show that genuine structure-response signal is present, but that the native-prompt accuracy depends substantially on pretrained, property-specific knowledge and familiar target scales. Overall, our results suggest a two-fold conclusion on the competitiveness of LLM-ICL: it succeeds in predicting chemical","source_metadata":{"date_field":"posted","keywords":[]}},{"id":"preprints:2611.25076v1","kind":"preprints","source":"arXiv","title":"ShEPhERD-score: Fast, adaptable three-dimensional alignment of molecular features","url":"https://arxiv.org/abs/2611.25076v1","detail_url":"/chemradar/article?u=https%3A%2F%2Farxiv.org%2Fabs%2F2611.25076v1","date":"2026-11-18T19:17:58Z","timestamp":1795029478,"categories":["Cheminformatics","Property Prediction","Docking & Screening"],"topic_ids":["cheminformatics","property","docking"],"keywords":["virtual screening","molecular features","molecular representations"],"matched_keywords":["virtual screening","machine learning","molecular features","molecular representations","molecular","ligand","protein"],"matched_tags":["virtual_screening","machine_learning","molecular_descriptor","generic_chem","biology"],"doi":null,"external_id":"2611.25076v1","pdf_url":"https://arxiv.org/pdf/2611.25076v1","code_url":"https://github.com/coleygroup/shepherd-score","code_host":"GitHub","authors":["Nate Somers","Kento A. Abeywardane","Connor W. Coley","Masha Elkin"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"Three-dimensional (3D) molecular similarity scoring and alignment are widely employed in ligand-based virtual screening and molecular design, and are increasingly used for training and evaluating machine learning models. Existing open-source methods utilize various molecular representations spanning shape, electrostatics, pharmacophores, and hybrid representations, but their implementations as independent software packages with distinct application programming interfaces (APIs), optimizers, and limited GPU support make them difficult to compare, combine, or extend. Here, we develop ShEPhERD-score into a unified, modular platform for differentiable 3D molecular similarity scoring and alignment by combining diverse molecular representations with high-performance Numba CPU and Triton GPU kernels to support both large-scale library screening and pairwise similarity workflows. In matched comparisons with open-source 3D similarity tools on the DUDE-Z virtual screening benchmark, ShEPhERD-score performed comparably or better in eight of nine cases. Across the scoring modes in ShEPhERD-score, throughput ranges from 17,800 to 1.03 million alignments/second for ten-million-conformer libraries using a single NVIDIA L40S GPU. Moreover, we found that the most effective scoring mode varied by protein target, indicating that target-specific calibration against validation data may be valuable before large-scale screening and underscoring the need for a unified platform that provides diverse molecular representations. Together, these capabilities establish ShEPhERD-score as a single, extensible toolkit for rapid virtual screening and machine learning workflows.","source_metadata":{"categories":["q-bio.BM"],"code_status":"found","code_url":"https://github.com/coleygroup/shepherd-score"}},{"id":"preprints:2611.25052v1","kind":"preprints","source":"arXiv","title":"MAPPA: Mapping of Atoms by Progressive Pattern Assembly with Molecular Symmetry","url":"https://arxiv.org/abs/2611.25052v1","detail_url":"/chemradar/article?u=https%3A%2F%2Farxiv.org%2Fabs%2F2611.25052v1","date":"2026-11-18T19:02:04Z","timestamp":1795028524,"categories":["Reaction Informatics"],"topic_ids":["reaction"],"keywords":["Atom to atom mapping"],"matched_keywords":["Atom-to-atom mapping","Molecular","chemical"],"matched_tags":["atom_mapping","generic_chem"],"doi":null,"external_id":"2611.25052v1","pdf_url":"https://arxiv.org/pdf/2611.25052v1","code_url":"https://github.com/yunhzou/MAPPA-AAM","code_host":"GitHub","authors":["Yunheng Zou","Olalla Nieto Faza","Shifa Hussain","Varinia Bernales","Alan Aspuru-Guzik"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"Atom-to-atom mapping traces atom identities between molecular endpoints, supporting reaction anal- ysis and pathway construction. Yet the same endpoints can admit multiple correspondences, and selecting a single answer can conceal alternatives with different chemical implications. We introduce MAPPA, a framework that efficiently discovers and retains alternative atom mappings. It uses shared molecular fragments and molecular symmetry to represent these alternatives compactly, allowing their bond changes to be examined and supporting subsequent chemistry-aware selection. MAPPA achieves the highest reference coverage among the evaluated methods on Golden. Standard benchmark runs re- cover 99.03% of Golden and 99.63% of EnzymeMap reference mappings, with matching and decoding together averaging 3.19 and 13.16 CPU seconds per reaction, respectively. Broader exploration raises Golden coverage to 99.51%. These results make broad exploration of mapping alternatives practical, providing a foundation for studying reaction ambiguity and incorporating chemical knowledge into mapping selection.","source_metadata":{"categories":["physics.chem-ph"],"code_status":"found","code_url":"https://github.com/yunhzou/MAPPA-AAM"}},{"id":"preprints:2611.25041v1","kind":"preprints","source":"arXiv","title":"Can Language Models Build Molecules in Three Dimensions?","url":"https://arxiv.org/abs/2611.25041v1","detail_url":"/chemradar/article?u=https%3A%2F%2Farxiv.org%2Fabs%2F2611.25041v1","date":"2026-11-18T19:00:08Z","timestamp":1795028408,"categories":[],"topic_ids":[],"keywords":["molecular shape"],"matched_keywords":["molecular shape","Molecules","Chemical","molecular","compounds","ligand"],"matched_tags":["shape_similarity","generic_chem"],"doi":null,"external_id":"2611.25041v1","pdf_url":"https://arxiv.org/pdf/2611.25041v1","code_url":"https://github.com/the-matter-lab/MolDraft-data","code_host":"GitHub","authors":["Hanyong Xu","Yunheng Zou","Olalla Nieto Faza","Juan Bernardo Pérez Sánchez","Aiwei Yin","Alán Aspuru-Guzik","Varinia Bernales"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"Chemical drawings convey which atoms are bonded and how stereocentres are configured, but not molecular shape. Every simulation starts from three-dimensional coordinates, and building them from a drawing remains a manual step. We introduce MolDraft, a curated benchmark of 300 drawing-structure pairs spanning natural products, hypervalent compounds, and coordination complexes. Given an image and a short instruction, an AI system must generate a complete set of atomic coordinates, which are checked for molecular identity, stereochemical configuration, and coarse ligand geometry. We evaluate nine multimodal language models using the same agent setup. The share of molecules built correctly, the resolve rate, ranges from 25.0% to 76.0% and falls as molecules grow, most sharply above about 80 atoms. Only 29 of the 300 molecules are resolved by all nine models, and literature structures, mostly metal complexes, are the hardest source for every model. Language models can thus build many molecules in three dimensions, but not yet reliably reproduce the coordination chemistry of the research literature.","source_metadata":{"categories":["physics.chem-ph"],"code_status":"found","code_url":"https://github.com/the-matter-lab/MolDraft-data"}},{"id":"preprints:2611.24764v1","kind":"preprints","source":"arXiv","title":"Horizontal Flow Matching for Single-Cell Spatiotemporal Dynamics","url":"https://arxiv.org/abs/2611.24764v1","detail_url":"/chemradar/article?u=https%3A%2F%2Farxiv.org%2Fabs%2F2611.24764v1","date":"2026-11-18T17:16:53Z","timestamp":1795022213,"categories":[],"topic_ids":[],"keywords":["transformer","equivariant"],"matched_keywords":["transformer","equivariant","molecular"],"matched_tags":["transformer","equivariance","generic_chem"],"doi":null,"external_id":"2611.24764v1","pdf_url":"https://arxiv.org/pdf/2611.24764v1","code_url":null,"code_host":null,"authors":["Zihan Wang","Jianzhe Li","Zhenyi Zhang","Peijie Zhou"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"Learning the spatiotemporal dynamics of cells from spatial transcriptomics requires connecting snapshots of different cells acquired in different coordinate frames. A rotation or translation of a section can appear as cell movement and distort the inferred correspondence. We introduce horizontal flow matching, which learns changes in tissue organization, molecular state, and cell abundance up to shared rigid motion. The dynamics follow a continuity--reaction equation whose action assigns no cost to shared rigid motion, and the resulting horizontal projector removes this component from spatial prediction errors during learning and from predicted velocities during generation. An equivalent endpoint problem jointly estimates correspondence and relative pose in Wasserstein--Fisher--Rao geometry and yields analytic paths whose velocity and growth serve as regression targets. Regressing these targets recovers the horizontal dynamics under ideal population conditions. Our model, HORIZON, uses an equivariant transformer to predict velocity and growth from the current tissue, so that a tissue can be generated continuously from a single initial snapshot. On a simulated tissue observed in perturbed coordinate frames, alignment improves reconstruction and projection reduces rigid drift. On four spatial transcriptomics time courses, HORIZON reconstructs tissue geometry, gene expression, and cell abundance more accurately than existing spatial and unbalanced trajectory methods, and interpolates held-out developmental stages.","source_metadata":{"categories":["cs.LG","cs.AI","q-bio.QM"],"code_status":"absent"}},{"id":"preprints:2611.24421v1","kind":"preprints","source":"arXiv","title":"Orbital-weighted learning for transition metal complexes","url":"https://arxiv.org/abs/2611.24421v1","detail_url":"/chemradar/article?u=https%3A%2F%2Farxiv.org%2Fabs%2F2611.24421v1","date":"2026-11-18T13:43:12Z","timestamp":1795009392,"categories":[],"topic_ids":[],"keywords":[],"matched_keywords":["neural network","machine learning","molecular","ligand"],"matched_tags":["neural_network","machine_learning","generic_chem"],"doi":null,"external_id":"2611.24421v1","pdf_url":"https://arxiv.org/pdf/2611.24421v1","code_url":null,"code_host":null,"authors":["Gabriel Klassen","Guido Falk von Rudorff"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"Frontier orbital energies are key to the description of molecular light-matter interaction and pose a challenge both for electronic structure calculations and for machine learning approaches, especially for transition metal complexes (TMC). These are highly relevant since the gap between the highest occupied molecular orbital (HOMO) and the lowest unoccupied molecular orbital (LUMO) is in the range of visible light. For TMCs, gap position and magnitude are determined by the metal center, its coordination and ligand identity. Since the systems exhibit a wide range of charge and spin states, frontier orbital energies can only be predicted with machine learning models that explicitly incorporate spin and charge. In this work, we show that a semi-empirical orbital population as a weight rather than a feature can increase efficiency for both physics-based and neural-network-based atomic representations. Because orbital populations are highly sensitive to both charge and spin, they offer a general pathway to render existing geometry-based machine learning representations useful for orbital energies. On TMC datasets, our approach reduces the required training data by up to 95 %, and improves physics-based representations to almost match the accuracy of pre-trained neural network representations at 12% of the inference compute cost. This greatly reduces the data and compute requirements to accurately model frontier orbitals in TMCs","source_metadata":{"categories":["physics.chem-ph"],"code_status":"absent"}},{"id":"preprints:2611.24148v1","kind":"preprints","source":"arXiv","title":"Machine learning interatomic potentials and orientational defects in polymers","url":"https://arxiv.org/abs/2611.24148v1","detail_url":"/chemradar/article?u=https%3A%2F%2Farxiv.org%2Fabs%2F2611.24148v1","date":"2026-11-18T09:54:37Z","timestamp":1794995677,"categories":["ML Potentials"],"topic_ids":["potentials"],"keywords":["MACE","density functional theory","DFT"],"matched_keywords":["Machine learning interatomic potentials","MACE","machine learning interatomic potential","Machine learning","density functional theory","DFT","molecular"],"matched_tags":["mlip_tools","machine_learning","dft","generic_chem"],"doi":"10.1007/978-94-024-2316-7_6","external_id":"2611.24148v1","pdf_url":"https://arxiv.org/pdf/2611.24148v1","code_url":null,"code_host":null,"authors":["Arnold Kiv","Volodymyr Soloviev","Andrii Tuzhykov","Andrii Bielinskyi","Taras Kavetskyy","Nataliia Hoivanovych","Ondrej Šauša","Jaroslav Rusnák","Helena Švajdlenková","Jolita Ostrauskaite","Mykola Slusarenko"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"Photopolymerization of acrylated epoxidized soybean oil (AESO) is a convenient model of a self-organizing process in a soft, bio-based material: ultraviolet light starts a free-radical chain reaction, the growing network releases heat, and the mobility of molecular groups changes on several time scales at once. We follow this process from two sides. Experimentally, the real and imaginary parts of the dielectric permittivity of an AESO:VDM (1:0.5) mixture with 3 mol% of the photoinitiator DMPA were recorded at 500-5000 Hz under an intermittent on/off illumination protocol, and the resulting time series were analysed with a sliding-window Shannon entropy and the Grassberger-Procaccia correlation integral. Both measures show that short light pulses at the start of curing trigger strongly correlated, collective relaxation, which fades as the network forms, leaving only localized reorientations active. We describe these reorientations as orientational defects and quantify their energetics with the MACE-OFF23 machine learning interatomic potential, using relaxed dihedral scans of five rotational modes ($α$, $α/β$, $β$, $γ$, $δ$) of an AESO fragment. Compared with density functional theory (DFT) reference profiles, MACE reproduces the shape of all five profiles (Pearson $r=0.95$-$0.98$) and the hierarchy of barriers - from 2.9 eV for the constrained main-chain rotation, through 0.54-0.81 eV for the acrylate rotations, to 0.16-0.18 eV for the methyl and 0.06 eV for the hydroxyl rotation - while underestimating the barrier heights by 3-20%. A simple Arrhenius estimate shows that the acrylate rotations are the modes closest to the frequency window of the dielectric measurement.","source_metadata":{"categories":["cond-mat.mtrl-sci","cond-mat.soft","physics.chem-ph"],"code_status":"absent"}},{"id":"preprints:2611.23873v1","kind":"preprints","source":"arXiv","title":"Technical Note on Matlantis PFP v9: Extending the Element and Domain Coverage of the r2SCAN Training Data","url":"https://arxiv.org/abs/2611.23873v1","detail_url":"/chemradar/article?u=https%3A%2F%2Farxiv.org%2Fabs%2F2611.23873v1","date":"2026-11-18T05:25:35Z","timestamp":1794979535,"categories":["ML Potentials"],"topic_ids":["potentials"],"keywords":["density functional theory","DFT","PBE"],"matched_keywords":["machine learning interatomic potentials","machine learning","density functional theory","DFT","PBE","molecules"],"matched_tags":["mlip_tools","machine_learning","dft","generic_chem"],"doi":null,"external_id":"2611.23873v1","pdf_url":"https://arxiv.org/pdf/2611.23873v1","code_url":null,"code_host":null,"authors":["Daiki Shintani","Shigeru Iwase","Jun Oshiki","Kohei Shinohara","Nontawat Charoenphakdee","Makoto Hiramatsu","Yuta Tsuboi","Yusuke Asano","Hirotaka Akita","Yong-Bin Zhuang","So Takamoto","Chikashi Shinagawa"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"Universal machine learning interatomic potentials (uMLIPs) make atomistic simulations possible at scales beyond the reach of density functional theory (DFT). Matlantis PFP is a uMLIP that treats a variety of domains, including crystals, molecules, and surfaces, with a single model. PFP v8 incorporated a training dataset computed with the r2SCAN functional, chosen for its good agreement with experiment, but the elements and domains covered by that dataset were limited. Here we present Matlantis PFP v9, in which the training data computed with the r2SCAN functional are extended to the lanthanides from Pr to Lu and to the elements from Po to Cm, for a total of 96 elements, and to additional domains such as clusters and adsorption structures. The formation energies of crystals containing the newly supported elements reproduce experimental standard enthalpies of formation with the same level of accuracy as crystals composed only of previously supported elements, and the adsorption energies of molecules on transition-metal surfaces have an overall error against experiment comparable to that of DFT-PBE, although chemisorbed species are overbound. On the same type of GPU (A100), the inference speed and the maximum system size exceed those of SevenNet-Omni, a public uMLIP trained on r2SCAN data: forces are evaluated three to four times faster, and five to ten times as many atoms can be handled. As a result, PFP-R2SCAN is now applicable to systems containing rare-earth and actinide elements and to adsorption on surfaces.","source_metadata":{"categories":["cond-mat.mtrl-sci","physics.comp-ph"],"code_status":"absent"}},{"id":"preprints:2611.23777v1","kind":"preprints","source":"arXiv","title":"STORM: A Lightweight Structure-Free Model for Binding Affinity Prediction","url":"https://arxiv.org/abs/2611.23777v1","detail_url":"/chemradar/article?u=https%3A%2F%2Farxiv.org%2Fabs%2F2611.23777v1","date":"2026-11-18T03:34:08Z","timestamp":1794972848,"categories":["Targets & Structures","Free Energy & MD"],"topic_ids":["targets","freeenergy"],"keywords":["Boltz 2","free energy perturbation","Binding Affinity"],"matched_keywords":["Boltz-2","free-energy perturbation","machine learning","protein","molecule","drug","Molecules","Binding Affinity"],"matched_tags":["ml_protein_ligand","free_energy_calcs","machine_learning","biology","generic_chem","bioactivity"],"doi":null,"external_id":"2611.23777v1","pdf_url":"https://arxiv.org/pdf/2611.23777v1","code_url":null,"code_host":null,"authors":["Ayush Pandit","Rohan V. Koodli","Wei Deng","Roxana Daneshjou","Ron O. Dror"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"Accurately predicting a small molecule's binding affinity to a protein, a critical task in drug discovery, remains challenging. Physics-based approaches like free-energy perturbation (FEP) have very high computational costs and require expert setup. Recent machine learning--based cofolding models have reported accuracy approaching that of FEP---and at least one of them, Boltz-2, has already been widely adopted for affinity prediction---but they remain expensive for large computational screens and even more expensive for retraining on newly collected experimental data. To address these challenges, we introduce STORM: Similarity Transfer from Observed Reference Molecules. STORM is a fast, simple machine learning model that ensembles many small model fusions of diverse protein and molecule representations for binding affinity prediction. Trained on the same data, STORM matches or exceeds the accuracy of Boltz-2 on multiple evaluation datasets, while delivering a speed improvement of over five orders of magnitude for both inference and training. Our results also suggest opportunities to substantially improve affinity prediction accuracy by integrating historical experimental affinity data with structural and biophysical information.","source_metadata":{"categories":["q-bio.QM","cs.LG"],"code_status":"absent"}},{"id":"journals:10.1039/d6dd00558f","kind":"journals","source":"Digital Discovery","title":"Pocket-Based 3D Molecular Generation: Binding-Environment Representations and Pocket Adaptability","url":"https://doi.org/10.1039/d6dd00558f","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1039%2Fd6dd00558f","date":"2026-11-18T02:43:43Z","timestamp":1794969823,"categories":["Design de novo"],"topic_ids":["denovo"],"keywords":["Molecular Generation"],"matched_keywords":["Molecular Generation","Molecular","drug","ligands","protein"],"matched_tags":["de_novo_design","generic_chem","biology"],"doi":"10.1039/d6dd00558f","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Bo Huang","Dongsheng Cao","Aiping Lu","Rong Qian","Tingfei Zhu","Jia-Cai Yi","Woruo Chen","Lei Wang","Yanjing Duan","Ling Zhang","Jin Liu"],"journal":"Digital Discovery","publisher":"RSC","impact_factor":5.6,"abstract":"Pocket-based 3D molecular generation has emerged as an important direction in structure-based drug design by enabling ligands to be generated directly within protein binding environments. Most existing methods rely on...","source_metadata":{"short_name":"Digital Discovery","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"preprints:10.26434/chemrxiv.15012428/v1","kind":"preprints","source":"ChemRxiv","title":"AIMETRY: GEOMETRY-SENSITIVE 3D MOLECULAR REPRESENTATIONS FROM QUANTUM MECHANICAL LEARNED FEATURES","url":"https://doi.org/10.26434/chemrxiv.15012428/v1","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.26434%2Fchemrxiv.15012428%2Fv1","date":"2026-11-18T00:00:00Z","timestamp":1794960000,"categories":["Cheminformatics"],"topic_ids":["cheminformatics"],"keywords":["ChEMBL","transformer","MOLECULAR REPRESENTATIONS"],"matched_keywords":["ChEMBL","machine learning","transformer","MOLECULAR REPRESENTATIONS","MOLECULAR","chemical","molecule","molecules"],"matched_tags":["chem_benchmarks","machine_learning","transformer","molecular_descriptor","generic_chem"],"doi":"10.26434/chemrxiv.15012428/v1","external_id":"10.26434/chemrxiv.15012428/v1","pdf_url":null,"code_url":null,"code_host":null,"authors":["Sheng Shu","Kaiya Taura","Shuhao Zhang","Olexandr Isayev"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"Quantifying molecular similarity depends on both chemical composition and 3D structure, which are difficult to capture in a single fixed-length representation. Contrastive learning is a common machine learning approach to generate such representations (molecular embeddings), pulling each molecule toward a positive anchor of itself. These anchors typically come from graph augmentations that alter chemical identity or from other conformers of the same molecule. However, using conformers as positive anchors in standard contrastive learning suppresses geometric distinctions and graded structural relationships. Here, we introduce AIMetry, a geometry-sensitive representation learning scheme in which a graph transformer aggregates spatially and chemically aware atomic features from AIMNet2, a machine-learned interatomic potential trained on quantum-chemical data, into a molecular embedding. We guide standard contrastive learning using fingerprint similarity, which modulates embedding similarity between different-molecule pairs, and root-mean-square deviation, which modulates embedding similarity between conformer pairs. Trained on approximately 44 million conformers of 2.8 million ChEMBL molecules, AIMetry demonstrates higher retention of fingerprint-defined structural relationships and greater geometric sensitivity than a reference model pretrained with standard contrastive learning. AIMetry also improves linear-probe performance relative to the reference model across the evaluated datasets, while benefits under nonlinear readout and fine-tuning are task-specific. Chemical and conformational analyses provide complementary evidence about the information retained in the representation.","source_metadata":{"date_field":"posted","keywords":[]}},{"id":"journals:10.1021/acs.jcim.6c02793","kind":"journals","source":"Journal of Chemical Information and Modeling","title":"Artificial Intelligence\nfor Generating Drug-like Molecules:\nA Taxonomy of Methods and Evaluation Strategies","url":"https://doi.org/10.1021/acs.jcim.6c02793","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1021%2Facs.jcim.6c02793","date":"2026-11-18T00:00:00+00:00","timestamp":1794960000,"categories":["Design de novo"],"topic_ids":["denovo"],"keywords":["molecule generation","generative molecular design","generative model"],"matched_keywords":["molecule generation","Generative artificial intelligence (AI) is increasingly used for de novo small-molecule design","generative molecular design","generative model","Drug","Molecules","molecule","molecular"],"matched_tags":["de_novo_design","generative_drug_design","generative_model","generic_chem"],"doi":"10.1021/acs.jcim.6c02793","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Chia-Chu Fan","Emily Wittrup","Morgen Henry","Gilbert S. Omenn","Kayvan Najarian"],"journal":"Journal of Chemical Information and Modeling","publisher":"ACS","impact_factor":5.3,"abstract":"Generative artificial intelligence (AI) is increasingly used for de novo small-molecule design, but the practical and translational utility of generated candidates depends on the configuration of the entire discovery workflow rather than on generative model architecture alone. This review adopts a workflow-based perspective on generative molecular design, tracing the process from existing molecular data sets, preprocessing and representation, to molecule generation, conditioning, candidate prioritization, evaluation, and iterative refinement. It examines how generative AI model families, constraints, and predictive oracles interact within this pipeline, and evaluates generated candidates using intrinsic metrics, computational assessments, and experimental and translational evidence. Active learning, computational and experimental feedback, and human expert input are further discussed as mechanisms for closed-loop refinement. Across these stages, we identify data construction and representation choices, oracle quality, evaluation protocols, and the availability of informative feedback as major determinants of how model performance should be interpreted and whether generated structures progress toward biologically and translationally meaningful candidates, with selected examples. In conclusion, reliable generative drug discovery requires coordinated design across the entire workflow rather than isolated advances in individual components.","source_metadata":{"short_name":"JCIM","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"journals:10.1021/acsomega.6c11449","kind":"journals","source":"ACS Omega","title":"Attenuated Total\nReflection Fourier Transform Infrared\nSpectroscopy and Chemometrics for the Forensic Investigation of Intentional\nRodenticide Poisoning in Ground and Liquid Coffee","url":"https://doi.org/10.1021/acsomega.6c11449","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1021%2Facsomega.6c11449","date":"2026-11-18T00:00:00+00:00","timestamp":1794960000,"categories":[],"topic_ids":[],"keywords":["s","chemometric"],"matched_keywords":["s","chemometric","Infrared Spectroscopy (ATR-FTIR) combined with chemometric","compounds"],"matched_tags":["chemometrics","analytical_chemistry_ai","generic_chem"],"doi":"10.1021/acsomega.6c11449","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Luana de O. Santos Simões","Gabriel M. Viegas","Victor G. P. Saide","José G. Rocha"],"journal":"ACS Omega","publisher":"ACS","impact_factor":null,"abstract":"Advances in forensic science and analytical technologies have significantly improved the collection, preservation, and interpretation of material evidence. Food and beverage matrices can serve as valuable forensic evidence when biological samples are absent or degraded. Among these matrices, coffee is particularly relevant to investigations of intentional poisoning because its sensory characteristics can mask toxic substances. Pesticides marketed as rodenticides, including carbamate and organophosphate compounds, together with the anticoagulant rodenticide brodifacoum, are of forensic interest because of their toxicity, widespread use, and occurrence in poisoning cases. Traditional chromatographic techniques, while robust and sensitive, are costly and labor-intensive, and many forensic laboratories face limited access to such equipment. Attenuated Total Reflection Fourier Transform Infrared Spectroscopy (ATR-FTIR) combined with chemometric tools such as Principal Component Analysis (PCA) and Soft Independent Modeling of Class Analogy (SIMCA) offers a rapid, nondestructive, and cost-effective alternative for screening rodenticide-related contamination in complex matrices. This study evaluated the potential of ATR-FTIR coupled with chemometric analysis for detecting rodenticide-related contamination in coffee and its applicability to forensic investigations of intentional poisoning. Commercial brodifacoum-based rodenticides and rodenticide formulations of unknown composition were investigated as representative toxic agents. A preliminary evaluation of ground coffee contaminated with one unknown formulation at 5.0 and 10.0% (w/w) was conducted to define the experimental conditions. The 5.0% level was selected for subsequent analyses because it was the lower of the two evaluated levels and already provided multivariate differentiation from uncontaminated samples. A total of 85 ground-coffee observations and 84 liquid-coffee observations were initially evaluated. Following Hotelling’s T2/Q-SPE assessment and inspection of the original spectral profiles, the refined data sets were used for exploratory and supervised multivariate analyses. Three SIMCA-based decision rules, data-driven SIMCA (DD-SIMCA), alternative SIMCA (Alt-SIMCA), and a Hotelling’s T2-only rule, were evaluated using the corrected spectral range for ground coffee or the full spectral range for liquid coffee, as well as the fingerprint region from 1800 to 900 cm–1, with external-validation sets comprising 26 to 28 samples depending on the matrix and spectral region. For ground coffee, both DD-SIMCA and Alt-SIMCA achieved 100% external-validation accuracy and 100% sensitivity for all classes in both spectral regions, with specificities of at least 75.00% and 80.00%, respectively. For liquid coffee, the best performance was obtained with Alt-SIMCA using the full spectral range, which achieved 100% external-validation accuracy, 100% sensitivity for all classes, and class specificities ranging from 89.47 to 100%. Within the tested formulations, coffee brands, and contamination level, these findings indicate the potential of ATR-FTIR combined with chemometrics as a rapid screening tool for the forensic analysis of coffee samples suspected of contamination with toxic agents.","source_metadata":{"short_name":"ACS Omega","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"journals:10.1021/acs.jcim.6c02429","kind":"journals","source":"Journal of Chemical Information and Modeling","title":"BALLROOM A and\nS–A Benchmark Created Automatically\nfrom Structurally Resolved Pockets with Homogeneously Measured Affinities/Series","url":"https://doi.org/10.1021/acs.jcim.6c02429","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1021%2Facs.jcim.6c02429","date":"2026-11-18T00:00:00+00:00","timestamp":1794960000,"categories":["Cheminformatics","Docking & Screening","Design de novo"],"topic_ids":["cheminformatics","docking","denovo"],"keywords":["ChEMBL","virtual screening","lead optimization","bioactivity"],"matched_keywords":["ChEMBL","virtual screening","protein","lead optimization","compound","molecule","ligand","molecular","bioactivity"],"matched_tags":["chem_benchmarks","virtual_screening","biology","lead_optimization","generic_chem","bioactivity"],"doi":"10.1021/acs.jcim.6c02429","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Torben Gutermuth","Tobias Harren","Matthias Rarey"],"journal":"Journal of Chemical Information and Modeling","publisher":"ACS","impact_factor":5.3,"abstract":"Computational methods for virtual screening and lead optimization are benchmarked on specifically designed data sets to predict compound potency. Here we present a workflow and two resulting new benchmark data sets, BALLROOM A and BALLROOM S, from combining PDB and ChEMBL data. Due to high data quality standards, the formation of molecule pairs measured under the same conditions only, and avoiding pairs with close to equal potency, we achieve a reliable data collection for physics-based bioactivity prediction based on protein–ligand complex structures. When testing state-of-the-art scoring functions for docking, no classical method statistically outperforms a simple molecular-weight-based prediction, consistent with previous studies such as D3R. While challenging for physics-based methods, the prediction task can be easily addressed by simple nearest-neigbor scorers showing that the benchmark cannot be used for machine learing based scoring functions due to significant train-test inference.","source_metadata":{"short_name":"JCIM","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"journals:10.1021/jacs.6c06587","kind":"journals","source":"Journal of the American Chemical Society","title":"De Novo\n                    Folding Mechanisms of Lasso\nPeptides","url":"https://doi.org/10.1021/jacs.6c06587","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1021%2Fjacs.6c06587","date":"2026-11-18T00:00:00+00:00","timestamp":1794960000,"categories":[],"topic_ids":[],"keywords":["Variational AutoEncoder","molecular dynamics"],"matched_keywords":["deep learning","Variational AutoEncoder","molecular","molecular dynamics"],"matched_tags":["machine_learning","vae","generic_chem","molecular_dynamics"],"doi":"10.1021/jacs.6c06587","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Song Yin","Xuenan Mi","Susanna E. Barrett","Mayuresh G. Gadgil","Douglas A. Mitchell","Diwakar Shukla"],"journal":"Journal of the American Chemical Society","publisher":"ACS","impact_factor":15.7,"abstract":"Lasso peptides adopt a distinctive [1] rotaxane conformation, yet the principles governing the folding of this kinetically trapped structure have remained elusive. Here, we integrated extensive molecular dynamics simulations and deep learning to elucidate the de novo folding mechanism of 20 lasso peptides lacking secondary post-translational modifications. We constructed Multi-Ensemble Markov Models for each lasso peptide and uncovered a consistent uphill folding landscape with spontaneous folding probabilities below 0.8%. Loop stability strongly correlated with folding propensity. Both simulations and experiments suggested that enhancing loop β-hairpin formation promotes folding of microcin J25, the well-studied lasso peptide extensively characterized as an in vitro model. This was confirmed by in silico loop-swap chimeras: grafting the microcin J25 loop onto the structurally similar klebsidin conferred prefolded stability. Leveraging Variational AutoEncoder-based pathway clustering, we resolved distinct pathway channels and representative folding pathways, further supporting this conclusion from a kinetic perspective. Additionally, a substantial entropy penalty was found to oppose lasso peptide folding. Simulations mimicking enzymatic spatial confinement reduced this penalty and stabilized the prefolded state. Together, these findings establish representative folding models and fundamental thermodynamic and kinetic principles which may be applied toward the rational engineering of lasso peptides.","source_metadata":{"short_name":"JACS","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"journals:10.1021/jacs.6c18843","kind":"journals","source":"Journal of the American Chemical Society","title":"De Novo Design of Single–Crystalline\nPeptide\nFrameworks Enables Structure–Encoded Molecular Recognition","url":"https://doi.org/10.1021/jacs.6c18843","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1021%2Fjacs.6c18843","date":"2026-11-18T00:00:00+00:00","timestamp":1794960000,"categories":[],"topic_ids":[],"keywords":["De Novo Design","Molecular dynamics"],"matched_keywords":["De Novo Design","Molecular","chemical","molecules","Molecular dynamics","protein"],"matched_tags":["de_novo_design","generic_chem","molecular_dynamics","biology"],"doi":"10.1021/jacs.6c18843","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Rui Yao","Jiachen Ye","Ying Liu","Liang Qiao","Yongsheng Wu","Zijuan Li","Bingyu Liu","Xiao Wang","Lifang Wu","Luofei Wei","Xiangkun Guo","Jie Liu","Lingmei Liu","Yong Cui","Jinqiao Dong"],"journal":"Journal of the American Chemical Society","publisher":"ACS","impact_factor":15.7,"abstract":"Molecular recognition by proteins underlies many biological processes and depends on the precise spatial organization of functional residues within well-defined binding pockets. Synthetic peptide assembly, however, is frequently biased toward anisotropic fibrillar or other low-dimensional morphologies, owing in part to the directional nature of peptide hydrogen-bonding networks, which can hinder the formation of ordered, structurally encoded binding cavities with biomimetic recognition properties. Here we show that coordination-driven β-sheet reprogramming overcomes this constraint by redirecting peptide assembly into single-crystalline peptide frameworks. Through de novo peptide design, we incorporate tyrosine, a residue that often plays a key role in protein molecular recognition, as a functional side chain, while metal–peptide connectivity suppresses fibrillar growth and decouples backbone packing from side-chain organization, thereby orienting tyrosine residues along internal channel surfaces to generate chiral recognition environments. The resulting frameworks exhibit high chemical and mechanical robustness, allow the spatial organization of tyrosine residues to be resolved crystallographically, and display excellent stereoselective recognition performance toward structurally complex glucocorticoids and chiral pharmaceuticals. Molecular dynamics simulations suggest that tyrosine residues dynamically interact with guest molecules during recognition. Our findings provide a design principle for constructing crystalline peptide frameworks with biomimetic molecular recognition, expanding the structural and functional scope of synthetic peptide materials.","source_metadata":{"short_name":"JACS","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"preprints:10.26434/chemrxiv.15012414/v1","kind":"preprints","source":"ChemRxiv","title":"Digital twin reconstruction of red blood cell aging using Raman chemometrics and systems modeling","url":"https://doi.org/10.26434/chemrxiv.15012414/v1","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.26434%2Fchemrxiv.15012414%2Fv1","date":"2026-11-18T00:00:00Z","timestamp":1794960000,"categories":["Spectra & Analytical"],"topic_ids":["chemometrics"],"keywords":["s","chemometric"],"matched_keywords":["s","chemometric"],"matched_tags":["chemometrics"],"doi":"10.26434/chemrxiv.15012414/v1","external_id":"10.26434/chemrxiv.15012414/v1","pdf_url":null,"code_url":null,"code_host":null,"authors":["Gabriela Sokorska","Anna Antolak","Aleksandra Wajda","Katarzyna Bulat","Natalia Wilkosz","Joanna Chwiej","Hugh J. Byrne","Katarzyna M. Marzec"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"Red blood cell (RBC) storage aging involves progressive metabolic and oxidative alterations that compromise membrane integrity and transfusion quality. However, conventional endpoint measurements provide limited information on individual aging dynamics and their predictability. This study presents an experimentally constrained digital twin framework integrating Raman spectral analysis, chemometric modeling, and mechanistic simulation to reconstruct RBC aging trajectories and predict donor-specific hemolysis. Longitudinal Raman spectra and biochemical measurements from packed RBC units stored for eight weeks were analyzed using principal component analysis, least squares decomposition, multivariate curve resolution–alternating least squares, and partial least squares regression. A MATLAB/SimBiology model coupled glucose metabolism, redox regulation, oxidative membrane deterioration, and hemoglobin release. The integrated approach resolved continuous biochemical trajectories characterized by glucose depletion, lactate accumulation, and progressive hemolysis. Despite broadly comparable metabolic kinetics, female-derived RBCs showed slower hemolysis than male-derived RBCs (0.05 ± 0.03 versus 0.40 ± 0.14 week-1). Early Raman spectral features predicted donor-specific hemolysis kinetics (R2 = 0.69), supporting their potential as predictive biomarkers of storage stability. The digital twin reproduced observed metabolic and hemolysis trajectories and implicated hydrogen peroxide regulation, glutathione availability, and GSH/GSSG turnover as candidate contributors to sex-dependent differences. By linking high-dimensional spectroscopic signals with interpretable mechanistic modeling, this framework provides a strategy for tracking RBC storage deterioration and estimating individual aging behavior from early measurements. It may support future development of personalized blood-quality assessment tools.","source_metadata":{"date_field":"posted","keywords":[]}},{"id":"preprints:10.26434/chemrxiv.15004164/v3","kind":"preprints","source":"ChemRxiv","title":"Distributed Assembly Theory: Distinguishing Biological Causal Depth from Information Entropy via Recursive Fragmentation","url":"https://doi.org/10.26434/chemrxiv.15004164/v3","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.26434%2Fchemrxiv.15004164%2Fv3","date":"2026-11-18T00:00:00Z","timestamp":1794960000,"categories":["Cheminformatics","Reaction Informatics"],"topic_ids":["cheminformatics","reaction"],"keywords":["SMILES","cheminformatics","retrosynthetic"],"matched_keywords":["SMILES","cheminformatics","molecular","molecules","retrosynthetic"],"matched_tags":["molecular_notation","cheminformatics","generic_chem","retrosynthesis"],"doi":"10.26434/chemrxiv.15004164/v3","external_id":"10.26434/chemrxiv.15004164/v3","pdf_url":null,"code_url":null,"code_host":null,"authors":["Chun Tang"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"Identifying agnostic biosignatures—measurable structural properties distinguishing evolved biochemistry from abiotic matter—remains a central challenge in astrobiology and cheminformatics [1, 13, 14]. Assembly Theory (AT) addresses this boundary through the molecular Assembly Index (M_A), a measure of constructional complexity defined by the minimal number of recursive bond-joining operations required to build a molecular graph while reusing previously assembled substructures [1, 15]. However, two major criticisms challenge this framework. First, critiques grounded in Algorithmic Information Theory (AIT) and Algorithmic Information Dynamics (AID) contend that M_A offers no fundamental advance over one-dimensional lossless compression, arguing that molecular assembly reduces to standard Lempel–Ziv dictionary proxies (zlib/DEFLATE and LZW) [4, 5, 7]. Second, computational critiques argue that exact M_A enumeration is intractable for complex natural products due to combinatorial subgraph explosion [1, 8]. Here, we resolve this methodological impasse. We demonstrate that while canonical branch-and-bound subgraph isomorphism logic [8] is mathematically sound, legacy internal parallelization layouts (specifically untied #pragma omp task directives using shallow firstprivate object copies) exhibit a 100% segmentation-fault crash rate on cyclic macromolecules due to race-condition double-free heap corruption. To eliminate this concurrency bottleneck, we introduce an asynchronous, process-isolated orchestration architecture paired with standardized MDL V2000 graph serialization that enforces strict single-threaded execution within individual C++ traversals while expanding compile-time bitset tracking bounds to 4,096 bits. High-throughput scaling is recovered via external operating-system process isolation (ProcessPoolExecutor) across a 256 GB unified-memory compute architecture. Using this decoupled pipeline, we evaluated 397 benchmark molecules across four structural complexity classes—achieving 396 exact integer resolutions (99.75%, with 1 combinatorial-horizon timeout at 1,200 s and zero mathematical bound violations) and resolving the diterpenoid Paclitaxel (Taxol) in isolation to M_A=23. For each resolved structure, we compared exact M_A against 1D dictionary compressibility quantified via zlib/DEFLATE compressed SMILES byte cost per non-hydrogen heavy atom (C_\"LZW\" =L_\"zlib\" (\"SMILES\" )/N_\"heavy\" ) through the Causal-to-Compression Ratio (R_\"MA/LZW\" =M_A/C_\"LZW\" ) [7, 12]. Whereas 1D byte cost per heavy atom inverts physical complexity due to syntactic character redundancy—scoring simple abiotic solvents higher than complex natural products (mean C_\"LZW\" =2.773 vs. 1.835\" bytes/heavy atom\" , with \"C\" _20 \"-\" \"C\" _40 branched alkanes lowest at 1.573)—both exact M_A and R_\"MA/LZW\" exhibit strictly monotonic divergence across complexity tiers. Mean R_\"MA/LZW\" climbs sequentially from abiotic solvents (2.656, mean M_A=4.83) and primary metabolites (3.926, mean M_A=7.90) through stochastic \"C\" _20 \"-\" \"C\" _40 branched alkanes (5.704, capped at a hard abiotic ceiling of M_A≤11) to complex natural products (7.214, mean M_A=12.48, benchmark maximum R_\"MA/LZW\" =16.580), reaching 15.333 (M_A=23 at C_\"LZW\" =1.50\" bytes/heavy atom\" ) for Paclitaxel. These results demonstrate that evolved biochemistry combines high 3D constructional depth with modular substructure reuse—a dual topological signature that 1D string compression fails to resolve. Note on Version History: This manuscript is a comprehensive revision of our preliminary preprint, \"Distributed Assembly Theory: Distinguishing Biological Causal Depth from Information Entropy via Recursive Fragmentation\" (ChemRxiv, June 2026) [9]. The preliminary version utilized a heuristic BRICS retrosynthetic fragmentation approximation [3], which violated exact ⌈log_2⁡E ⌉ graph lower bounds on fused polycyclic macromolecules. This V2 manuscript supersedes the preliminary report by implementing exact C++ branch-and-bound graph enumeration [8], resolving OpenMP heap-corruption vulnerabilities, and establishing exact M_A/C_\"LZW\" divergence baselines.","source_metadata":{"date_field":"posted","keywords":[]}},{"id":"preprints:10.26434/chemrxiv.15012418/v1","kind":"preprints","source":"ChemRxiv","title":"Drug-resistance phenotypes in FGFR deep mutational scanning are carried by a single-residue positional statistic that transfers across FGFR1–3 and weakens in FGFR4","url":"https://doi.org/10.26434/chemrxiv.15012418/v1","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.26434%2Fchemrxiv.15012418%2Fv1","date":"2026-11-18T00:00:00Z","timestamp":1794960000,"categories":[],"topic_ids":[],"keywords":["AutoDock Vina","kinase"],"matched_keywords":["AutoDock Vina","Drug","drugs","kinase"],"matched_tags":["docking_software","generic_chem","biology"],"doi":"10.26434/chemrxiv.15012418/v1","external_id":"10.26434/chemrxiv.15012418/v1","pdf_url":null,"code_url":null,"code_host":null,"authors":["Jia He"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"Motivation: Drug-resistance mutations are a clinically central but poorly predicted phenotype: pathogenicity predictors trained on disease labels misrank them, and our recent benchmark of five static predictor families (AlphaMissense, FoldX stability, FoldX interaction energy, ThermoMPNN-D and AutoDock Vina) showed that biophysical scores predict FGFR resistance poorly. Whether supervised recalibration on functional deep mutational scanning (DMS) labels can repair this, and what feature class carries the signal, remains untested. Results: Using the DMS of the FGFR1–4 kinase domains (Tangermann et al. 2026; 7,522 unique variants) under leave-one-subtype-out validation, raw AlphaMissense correlates negatively with the resistance score in every subtype (ρ = −0.46 to −0.65), and on the binary task models built on biophysical scores reach AUROC 0.54–0.74 and never exceed the position-based model (scores 0.56–0.64 vs position 0.73–0.87 on the complete-case subset, n = 5,135). On the continuous score, a one-parameter sign flip remains the strongest recalibration. The signal is instead carried by an aligned-position prior: a parameter-free lookup of resistance rates at each aligned kinase-domain coordinate, transferred from the training subtypes. The lookup matches or exceeds a gradient-boosted model (AUROC 0.88, 0.89, 0.84, 0.70 for FGFR1–4), is strongly resistance-skewed (pure-resistance 0.86, 0.84, 0.84, 0.61; pure-activating 0.62, 0.60, 0.42, 0.46), peaks at single-residue resolution, sits 0.16–0.30 above the strongest clinical-hotspot baseline, and reaches AUPRC 0.31–0.48 with 37–63% precision in the top 50; a two-stage blend that resolves within-site ordering raises this to 36–64% with a tie-free ranking at no AUROC cost. The prior transfers across drugs (pemigatinib → futibatinib, 0.74–0.83) and across FGFR1–3, weakening for FGFR4. Conclusion: FGFR resistance phenotypes are recoverable from a homology-transferred positional label statistic — a lookup table, not a fitted model; the lookup alone recovers the discrimination, and a within-site score ranker refines the top of the list without adding discrimination. The statistic is specific to resistance, contains information beyond known clinical hotspots, and transfers across drugs within the FGFR inhibitor class.","source_metadata":{"date_field":"posted","keywords":[]}},{"id":"journals:10.1021/acs.jcim.6c02995","kind":"journals","source":"Journal of Chemical Information and Modeling","title":"ezABFE: An Easy-to-Use\nWorkflow for Absolute Binding\nFree Energy Calculations","url":"https://doi.org/10.1021/acs.jcim.6c02995","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1021%2Facs.jcim.6c02995","date":"2026-11-18T00:00:00+00:00","timestamp":1794960000,"categories":["Free Energy & MD"],"topic_ids":["freeenergy"],"keywords":["ABFE","Free Energy Calculations","binding free energy","AMBER","receptor"],"matched_keywords":["ABFE","Free Energy Calculations","binding free energy","AMBER","ligands","ligand","receptor"],"matched_tags":["free_energy_calcs","thermodynamics","force_fields","generic_chem","biology"],"doi":"10.1021/acs.jcim.6c02995","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Xin Li","Yuanhao Li","Shipeng Wang","Jiake Li","Zhicheng Zuo"],"journal":"Journal of Chemical Information and Modeling","publisher":"ACS","impact_factor":5.3,"abstract":"Absolute binding free energy (ABFE) calculations can evaluate chemically diverse ligands, but their complex workflows remain challenging for nonexpert users and prone to errors. Here, we present ezABFE, an easy-to-use Python workflow for automated ABFE calculations within the AMBER ecosystem. ezABFE minimizes manual intervention in system preparation, representative-structure selection, Boresch restraint atom selection, simulation setup, free-energy estimation, and graphical analysis. The workflow employs both MBAR and TI for alchemical free-energy estimation and additionally provides MM-GBSA and optional MM-PBSA end-point calculations. Benchmarking across T4 lysozyme L99A, BRD4(1), CDK8, and the theophylline-binding RNA aptamer shows good overall agreement with experimentally derived binding free energies and comparable performance of MBAR and TI. Additional tests demonstrate that appropriate Boresch restraint atom selection is important for simulation stability and binding free-energy accuracy, and that ligand partial-charge models and the HMR/4 fs simulation protocol can influence calculated affinities, although the limited comparisons do not support general conclusions regarding either factor. The computationally inexpensive end-point methods also retain practical value for ligand ranking. Overall, ezABFE provides an accessible, automated workflow for reproducible ABFE calculations with improved throughput across diverse receptor systems, including nucleic acid receptors.","source_metadata":{"short_name":"JCIM","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"journals:10.1088/2632-2153/aec40f","kind":"journals","source":"Machine Learning: Science and Technology","title":"Fast contracted Clebsch-Gordan tensor products for equivariant graph neural networks","url":"https://doi.org/10.1088/2632-2153/aec40f","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1088%2F2632-2153%2Faec40f","date":"2026-11-18T00:00:00+00:00","timestamp":1794960000,"categories":[],"topic_ids":[],"keywords":["machine learning potentials","graph neural networks","message passing","equivariant"],"matched_keywords":["machine learning potentials","neural networks","graph neural networks","message passing","machine learning","equivariant"],"matched_tags":["mlip_tools","neural_network","gnn","machine_learning","equivariance"],"doi":"10.1088/2632-2153/aec40f","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Anton Bochkarev","Yury Vyatcheslavovich Lysogorskiy","Ralf Drautz"],"journal":"Machine Learning: Science and Technology","publisher":"IOP","impact_factor":6.3,"abstract":"We present an algorithm with an O(L3) operation count for evaluating contracted Clebsch–Gordan tensor products in O(3)-equivariant machine learning potentials at fixed Canonical Polyadic (CP) rank. Mapping the angular integral to a structured Gauss–Legendre and Fourier tensor-product grid decouples the radial channel contractions from the angular transforms. The antisymmetric parity- odd Clebsch–Gordan channels, unreachable by the symmetric pointwise product on a scalar S2 grid, are recovered through the surface-curl pairing ˆ r·[∇S2 A×∇S2 B], the spherical Poisson bracket, which supplies the L = 1 angular momentum on the grid while preserving rotational equivariance. The construction extends to parity-aware equivariant message passing in architectures of the atomic cluster expansion family and is verified by direct numerical quadrature. The full uncontracted Clebsch–Gordan tensor product remains subject to the O(L4) output-size lower bound. Benchmarks show wall-clock scaling approximately as L2 for both operations over the whole range measured, with the runtime on the tested GPU set by data movement on the L2-sized grid tensors rather than by the L3 arithmetic.","source_metadata":{"short_name":"MLST","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"journals:10.1186/s13321-026-01328-w","kind":"journals","source":"Journal of Cheminformatics","title":"Feature selection-guided graph neural networks for compound activity prediction","url":"https://doi.org/10.1186/s13321-026-01328-w","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1186%2Fs13321-026-01328-w","date":"2026-11-18T00:00:00+00:00","timestamp":1794960000,"categories":["Cheminformatics","Property Prediction"],"topic_ids":["cheminformatics","property"],"keywords":["graph neural networks","GNNs","GNN","message passing","activity prediction","property prediction","molecular descriptors","molecular property"],"matched_keywords":["graph neural networks for compound activity prediction Abstract Molecular graphs","graph neural networks (GNNs) for compound classification and regression tasks. While chemical information captured by molecular graphs","molecular graphs as input for GNNs","neural networks","graph neural networks","GNNs","GNN","message passing","activity prediction","property prediction","molecular descriptors","compound","Molecular","chemical","molecular property"],"matched_tags":["molecular_graph_method","neural_network","gnn","ml_task","molecular_descriptor","generic_chem","bioactivity"],"doi":"10.1186/s13321-026-01328-w","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Fouad Alkhoury","Tamás Horváth","Tiago Janela","Andrea Mastropietro","Jürgen Bajorath"],"journal":"Journal of Cheminformatics","publisher":"Springer","impact_factor":5.7,"abstract":"Molecular graphs serve as input for graph neural networks (GNNs) for compound classification and regression tasks. While chemical information captured by molecular graphs is often sufficient for compound classification and molecular property prediction tasks, it is often attempted to further increase the chemical information content of graph representations by augmenting them with various molecular descriptors. For this purpose, several approaches have been introduced that operate at different stages of GNN training, either before, during, or after message passing. Despite their methodological differences, these approaches commonly challenge GNNs to reduce feature noise, redundancy, or biases during learning, which often negatively affects predictions. As a conceptual alternative, we introduce a two-stage methodology in which a feature ranking GNN is used first to automatically select features important for compound activity prediction, without requiring the calculation of any descriptor values at this stage, followed by the derivation of a classification GNN based on the preselected feature set. For different compound activity classes and other datasets, we show that small preselected descriptor subsets achieve GNN predictive performance comparable to or better than original descriptor pools used for graph augmentation and also comparable to reference methods. Control calculations reveal that small preselected descriptor subsets add task-specific information for molecular graph classification, consistently outperforming randomly chosen subsets of equal size. Furthermore, we demonstrate that selected features aid in explaining predictions, are chemically interpretable, and represent activity determinants. Scientific contribution We present a new methodological concept for increasing the chemical information content of molecular graphs as input for GNNs through feature augmentation. The approach separates the augmentation and learning phases by using a feature ranking GNN in combination with a classification GNN. This approach essentially maintains or increases predictive performance and supports the explanation of predictions.","source_metadata":{"short_name":"J. Cheminform.","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"journals:10.1021/acsomega.6c04071","kind":"journals","source":"ACS Omega","title":"Fingerprinting, Chemical Constituents, and Quality\nControl of\n                    Dracocephalum tanguticum\n                    Maxim. Based\non HPLC and UPLC-MS/MS","url":"https://doi.org/10.1021/acsomega.6c04071","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1021%2Facsomega.6c04071","date":"2026-11-18T00:00:00+00:00","timestamp":1794960000,"categories":["Spectra & Analytical"],"topic_ids":["chemometrics"],"keywords":["chemometric"],"matched_keywords":["chemometric","chromatography coupled with Orbitrap high-resolution tandem mass spectrometry (UPLC-MS/MS), fingerprinting, and chemometric","Chemical","compounds"],"matched_tags":["chemometrics","analytical_chemistry_ai","generic_chem"],"doi":"10.1021/acsomega.6c04071","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Tingting Wang","Baoxin Zhang","Yongheng Li","Zhaoyue Dong","Yongsheng Wei","Hong Quan","Min Chen","Xiaozhong Lan"],"journal":"ACS Omega","publisher":"ACS","impact_factor":null,"abstract":"This study employed high-performance liquid chromatography (HPLC), ultraperformance liquid chromatography coupled with Orbitrap high-resolution tandem mass spectrometry (UPLC-MS/MS), fingerprinting, and chemometric methods to systematically analyze 37 batches of wild Dracocephalum tanguticum Maxim. samples collected from Xizang. A characteristic HPLC fingerprint was established, and UPLC-MS/MS was used for the rapid identification of chemical constituents. Potential quality markers were preliminarily selected based on high-resolution mass spectrometry data and literature review, followed by quantification across all samples. Principal component analysis (PCA) and hierarchical cluster analysis (HCA) were conducted using R to evaluate the quality variation among different origins based on the contents of selected quality markers. The results showed that 32 common peaks were extracted from the 37 batches, with considerable variability in the similarity among samples. A total of 53 compounds were identified, and 9 quality markers were ultimately screened through comparison with reference standards and MS interpretation. PCA based on these nine markers revealed that the first two principal components (PC1 and PC2) explained 53.3% of the total variance, effectively capturing the structural differences among the samples. Furthermore, HCA heatmaps combined with K-means clustering results demonstrated that the 37 batches could be grouped into three major clusters, which were highly consistent with PCA-based classification, confirming the reliability of the clustering. These findings provide a theoretical and data-based foundation for quality evaluation, grading control, and origin traceability of D. tanguticum, offering support for quality assurance in food and pharmaceutical industries.","source_metadata":{"short_name":"ACS Omega","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"preprints:10.26434/chemrxiv.15012459/v1","kind":"preprints","source":"ChemRxiv","title":"FOdorant conformational ensembles encode topology, molecular size and polar-surface coherence, not shape","url":"https://doi.org/10.26434/chemrxiv.15012459/v1","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.26434%2Fchemrxiv.15012459%2Fv1","date":"2026-11-18T00:00:00Z","timestamp":1794960000,"categories":["Property Prediction"],"topic_ids":["property"],"keywords":["Graph Neural Network","molecular representation","receptor"],"matched_keywords":["Neural Network","Graph Neural Network","molecular representation","molecular","receptor"],"matched_tags":["neural_network","gnn","molecular_descriptor","generic_chem","biology"],"doi":"10.26434/chemrxiv.15012459/v1","external_id":"10.26434/chemrxiv.15012459/v1","pdf_url":null,"code_url":null,"code_host":null,"authors":["Yuan SHANG","Yuan Shang"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"Odor is widely modelled as three-dimensional shape recognition: a flexible odorant presents a 300 K conformer ensemble to a receptor repertoire, implying that perceptual breadth peaks at an intermediate number of accessible shapes and that explicit 3D representations outperform 2D molecular graphs. We test both predictions under a cryptographically pre-registered plan across catalogue profiles (n = 4,808), human psychophysics (n = 304) and heterologous receptor arrays (n = 294). Computing 300 K GFN2-xTB ensembles at a 200-conformer ceiling for all 1,821 flexible cohort members (106,526 conformers) and benchmarking 21 neural representations across 10 seeds, we find that neither claim holds. Conformational entropy saturates by 28 conformers (88% attained at a ceiling of 20), excluding an interior breadth optimum sixfold below the populated range (β2 = +0.0080, 95% CI -0.0014 to +0.0172). No 3D representation outperforms a capacity-matched or descriptor-augmented 2D graph. Out-of-sample-replicated confound analyses show that 3D shape dispersion vanishes once chain topology is controlled; instead, breadth reflects coarse topological flexibility (peaking at 9.2 rotatable bonds), molecular extent and polar-surface coherence (concave β2 = -0.0264). Keywords: conformational ensemble · odor breadth · molecular representation · Artificial Intelligence · Graph Neural Network","source_metadata":{"date_field":"posted","keywords":[]}},{"id":"preprints:10.26434/chemrxiv.15012413/v1","kind":"preprints","source":"ChemRxiv","title":"From De Novo Molecular Design to Experimentally Validated Protein-Protein Interaction Modulators","url":"https://doi.org/10.26434/chemrxiv.15012413/v1","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.26434%2Fchemrxiv.15012413%2Fv1","date":"2026-11-18T00:00:00Z","timestamp":1794960000,"categories":["Design de novo"],"topic_ids":["denovo"],"keywords":["De Novo Molecular Design","molecular generation","generative molecular design","pharmacokinetic","ADMET","receptor"],"matched_keywords":["De Novo Molecular Design","molecular generation","generative molecular design","pharmacokinetic","ADMET","Molecular","molecules","drug","molecule","compounds","chemical","ligand","ligands","Protein","receptor"],"matched_tags":["de_novo_design","generative_drug_design","admet","generic_chem","biology"],"doi":"10.26434/chemrxiv.15012413/v1","external_id":"10.26434/chemrxiv.15012413/v1","pdf_url":null,"code_url":null,"code_host":null,"authors":["Óscar Toledano","Celia Miró-Rodríguez","Lorena Aguado","Eugenia Ulzurrun","Paula Alonso-García","Alicia Mansilla","María José Sánchez Barrena","Nuria E. Campillo"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"Identifying bioactive small molecules requires balancing target affinity with physicochemical and pharmacokinetic properties compatible with drug development. This is particularly challenging in the development of small-molecule modulators of protein–protein interactions (PPIs) due to the structural complexity of protein interfaces. Here, we developed an integrated workflow combining de novo molecular design with similarity-guided identification of existing compounds from public chemical libraries. We applied this strategy to neuronal calcium sensor-1 (NCS-1), a calcium-binding protein whose interactions with multiple partners regulate synaptic function, making it an attractive target for modulating neuronal signaling. REINVENT4-based molecular generation incorporated predicted binding to the NCS-1 hydrophobic crevice, synthetic accessibility, and CNS-relevant ADMET properties, including blood–brain barrier permeability. Generated structures guided the selection of commercially available analogues. Experimental validation supported the effectiveness of this AI-guided strategy: 11 of the 12 tested compounds bound NCS-1, including 5 with micromolar affinities. In situ protein–protein interaction assays further confirmed the modulatory activity of selected compounds, identifying, to our knowledge, the first small-molecule enhancers of the NCS-1/cannabinoid receptor 1 (CB1R) interaction. CB1R is an attractive pharmacological target implicated in neuronal signaling, whose activity is regulated by NCS-1. High-resolution crystal structures revealed how these modulators engage the NCS-1 hydrophobic crevice, providing structural insights into ligand recognition and suggesting a mechanism of action. These findings connect generative molecular design with experimentally accessible NCS-1 ligands and provide chemical tools and structural insights for investigating NCS-1-dependent regulation of CB1R.","source_metadata":{"date_field":"posted","keywords":[]}},{"id":"journals:10.1021/jacs.6c14994","kind":"journals","source":"Journal of the American Chemical Society","title":"High-Throughput\nScreening of Photocatalytic Metal–Organic\nFrameworks for CO2 Reduction Using Relational Data Extraction\nand Hierarchical Machine Learning","url":"https://doi.org/10.1021/jacs.6c14994","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1021%2Fjacs.6c14994","date":"2026-11-18T00:00:00+00:00","timestamp":1794960000,"categories":["LLMs & Agents"],"topic_ids":["llm"],"keywords":["convolutional neural","LLM","property prediction","reaction conditions"],"matched_keywords":["convolutional neural","LLM","Machine Learning","property prediction","reaction conditions"],"matched_tags":["neural_network","LLMs","machine_learning","ml_task","reaction_conditions"],"doi":"10.1021/jacs.6c14994","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Xiao Niu","Yan Liu","Yong Cui","Jianwen Jiang"],"journal":"Journal of the American Chemical Society","publisher":"ACS","impact_factor":15.7,"abstract":"Photocatalytic carbon dioxide reduction (CO2RR) over metal–organic frameworks (MOFs) is governed by both framework features and reaction conditions, which complicate catalyst discovery and render general design rules difficult to establish. In this study, we develop a synergistic machine-learning workflow to integrate literature data extraction, electronic-property prediction, and hierarchical classification for prioritizing MOF photocatalysts. First, a literature-derived record-level data set containing catalyst compositions, structural descriptors, reaction conditions, and catalytic outcomes was curated through large language model (LLM)-assisted relation extraction followed by manual verification. Then, a transfer-learning crystal graph convolutional neural network (CGCNN) model was trained for band gap prediction at the Heyd-Scuseria-Ernzerhof screened hybrid functional (HSE06) level, achieving a mean absolute error (MAE) of 0.31 eV against reference HSE06 calculations. Subsequently, a physically constrained multitask framework was employed to predict electron affinity (EA) and ionization potential (IP) for thermodynamic prescreening. Three record-level classifiers were constructed for activity screening (C1), performance classification (C2), and product selectivity prediction (C3). When applied to the Computation-Ready, Experimental (CoRE) MOF database under standardized gas- and liquid-phase inference settings, the workflow prioritized 639 primary candidates, with most assigned to CO formation. Probability calibration and applicability-domain analysis further identified 166 of these candidates (25.98%) as a more stringent calibrated, in-domain subset. Overall, this study provides a practical data-driven workflow for condition-resolved prioritization of MOF photocatalysts and analysis of descriptor-level trends associated with photocatalytic CO2RR.","source_metadata":{"short_name":"JACS","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"journals:10.1021/acsomega.6c10660","kind":"journals","source":"ACS Omega","title":"How Far Do Kamlet–Taft Parameters and pH Predict\nDelignification by Deep Eutectic Solvents?","url":"https://doi.org/10.1021/acsomega.6c10660","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1021%2Facsomega.6c10660","date":"2026-11-18T00:00:00+00:00","timestamp":1794960000,"categories":[],"topic_ids":[],"keywords":["random forest"],"matched_keywords":["random forest","chemical"],"matched_tags":["ensemble","generic_chem"],"doi":"10.1021/acsomega.6c10660","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Guadalupe Pérez-Durán","Supti Paul","Swati Sood","Alberto Gutierrez-Vega","Santiago Aparicio","Gustavo A. Iglesias-Silva","Roman Brunecky","Kecheng Li","Mert Atilhan"],"journal":"ACS Omega","publisher":"ACS","impact_factor":null,"abstract":"Deep eutectic solvents (DESs) are tunable, low-toxicity media for lignocellulose fractionation, but their selection for delignification still relies on empirical screening. Kamlet–Taft (KT) solvatochromic parameters (hydrogen-bond acidity α, basicity β, and dipolarity/polarizability π*) and pH are the descriptors most often correlated with lignin solubility, yet how far they predict delignification of whole biomass has not been tested critically. Here we report measured KT parameters and apparent pH (pHapp) for 56 DES formulations spanning carboxylic-acid, polyol, metal-salt, amide, alkanolamine, and amino-acid families, together with the softwood lignin removal obtained after a sequential enzymatic–DES fractionation. Four descriptor-to-performance models were compared under identical leave-one-out, repeated k-fold, and grouped (leave-one-cluster-out and leave-one-family-out) cross-validation: a linear correlation, a 15-term second-order polynomial, a random forest (RF), and a Gaussian-process (GP) regressor with automatic relevance determination. All models fit the training set (in-sample R2 up to 0.89 for the RF) but generalize weakly: leave-one-out R2 is 0.10–0.13 for the GP and the RF, the polynomial is overparameterized and generalizes worse than the training mean, and no model predicts a chemical family that is absent from training better than the training mean. Gaussian-process relevance and ablation and random-forest permutation importance rank pHapp as the most informative descriptor, followed by α, whereas β and π* carry little consistent information, and the pKa of the hydrogen-bond donor carries the same information as pHapp; these results point to solvent acid strength, rather than hydrogen-bond donation alone, as the property most closely associated with lignin removal from enzymatically pretreated softwood. The GP provides calibrated predictive uncertainty: 96% of leave-one-out predictions and all four prospectively synthesized blind-test DESs fall within its 95% credible intervals. The results delimit what solvatochromic descriptors can and cannot predict for whole-biomass delignification and provide an uncertainty-aware screening tool and an open data set for DES design.","source_metadata":{"short_name":"ACS Omega","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"journals:10.1021/acsomega.6c04840","kind":"journals","source":"ACS Omega","title":"In Silico Prediction\nof Bacterial Reverse Mutagenicity\nin Emerging Synthetic Cannabinoids: Implications for Early Carcinogenic\nRisk Assessment","url":"https://doi.org/10.1021/acsomega.6c04840","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1021%2Facsomega.6c04840","date":"2026-11-18T00:00:00+00:00","timestamp":1794960000,"categories":["Property Prediction","ADMET & Safety"],"topic_ids":["property","admet"],"keywords":["QSAR"],"matched_keywords":["QSAR","compound","Compounds","chemical","molecular"],"matched_tags":["qsar_qspr","generic_chem"],"doi":"10.1021/acsomega.6c04840","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Marcos Messias","Rafael Lanaro","Alexandre Barcia Godoi","Jose Luiz Costa"],"journal":"ACS Omega","publisher":"ACS","impact_factor":null,"abstract":"Synthetic cannabinoids (SCs) are a structurally diverse class of new psychoactive substances with scarce toxicological data and rapidly evolving structures that challenge risk assessment. This study applied the CASE Ultra, version 1.9.0.8 expert system to predict the genotoxic potential of 18 SCs identified in Brazilian seizures and the literature, and to evaluate how subtle structural modifications may influence mutagenicity-related alerts. Predictions were generated combining the GT_EXPERT (rule-based), GT1_BMUT (statistical bacterial reverse mutagenicity model), and the integrated Konsolidator module. For each compound, the software provided positive, negative, or inconclusive outputs, identifying structural alerts and deactivating fragments, as well as listed database analogs. Inconclusive or conflicting results underwent expert review considering structural similarity, DNA accessibility, metabolic plausibility, and supporting evidence. Compounds showing positive or inconclusive outcomes were further evaluated using three complementary QSAR platforms (VEGA Hub version 1.1.5, Toxtree 3.1.0, and OECD QSAR Toolbox version 4.8) to assess the consistency of structural alerts across different computational approaches. Whenever available, predictions were interpreted in light of published experimental data. Most SCs were classified as negative by the integrated CASE Ultra prediction, suggesting an overall low predicted bacterial reverse mutagenicity risk within the evaluated group. JWH-018 was the only compound consistently classified as positive, whereas CH-FUBIATA yielded an inconclusive prediction due to the coexistence of activating and deactivating structural features. Comparative analysis demonstrated good agreement between CASE Ultra and VEGA Hub, with both platforms identifying similar mutagenicity-associated structural alerts for JWH-018 and CH-FUBIATA. In contrast, neither Toxtree nor the OECD QSAR Toolbox identified structural alerts for these compounds, likely reflecting the limited representation of contemporary SCs within their underlying knowledge bases. A simulated structural modification of PX-2 demonstrated that removal of a deactivating amine group converted a negative prediction into a positive one, illustrating how minor chemical changes may substantially alter predicted genotoxic risk. Conversely, the predicted metabolites of PX-1 and PX-2 retained negative classifications, indicating that molecular modifications do not inherently alter mutagenic potential and that distinct structural changes, whether resulting from deliberate chemical modification or metabolic biotransformation, may lead to different genotoxicity outcomes. These findings demonstrate that predicted bacterial reverse mutagenicity is determined by the interaction between activating and deactivating structural features within the overall molecular architecture rather than by isolated pharmacophoric elements. The concordance observed between CASE Ultra and VEGA Hub supports the mechanistic interpretation of the identified structural alerts, whereas the absence of alerts in other expert systems underscores the need to expand QSAR training data sets for emerging substances. Overall, integrated in silico approaches represent valuable tools for the early prioritization of toxicological risks, although further experimental validation remains essential to strengthen predictive confidence.","source_metadata":{"short_name":"ACS Omega","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"feeds:https://chem-bla-ics.linkedchemistry.info/2026/11/18/interdisciplenary-or-transdisciplenary.html","kind":"feeds","source":"chem-bla-ics (Linked Chemistry)","title":"Interdisciplenary or transdisciplenary?","url":"https://chem-bla-ics.linkedchemistry.info/2026/11/18/interdisciplenary-or-transdisciplenary.html","detail_url":"/chemradar/article?u=https%3A%2F%2Fchem-bla-ics.linkedchemistry.info%2F2026%2F11%2F18%2Finterdisciplenary-or-transdisciplenary.html","date":"2026-11-18T00:00:00+00:00","timestamp":1794960000,"categories":["Blog"],"topic_ids":[],"keywords":[],"matched_keywords":[],"matched_tags":[],"doi":null,"external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":[],"journal":null,"publisher":null,"impact_factor":null,"abstract":"","source_metadata":{"source":"chem-bla-ics (Linked Chemistry)","published_utc":"2026-11-18T00:00:00+00:00","seen_at":"2026-11-18T14:21:00.676104+00:00"}},{"id":"journals:10.1021/acs.jctc.6c01216","kind":"journals","source":"Journal of Chemical Theory and Computation","title":"LIG-PROSPECT: Sequence-Transferable\nMachine Learning\nPrediction of Flavin-Dependent Photoenzyme Spectral Properties","url":"https://doi.org/10.1021/acs.jctc.6c01216","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1021%2Facs.jctc.6c01216","date":"2026-11-18T00:00:00+00:00","timestamp":1794960000,"categories":[],"topic_ids":[],"keywords":[],"matched_keywords":["Machine Learning","chemical","LIGand","PROtein"],"matched_tags":["machine_learning","generic_chem","biology"],"doi":"10.1021/acs.jctc.6c01216","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Bhumika Jayee","Sunny Lee","Sijia S. Dong"],"journal":"Journal of Chemical Theory and Computation","publisher":"ACS","impact_factor":null,"abstract":"Predicting the optical properties of proteins is critical for the understanding and design of photoactive proteins. Red-shifting flavin-dependent photoenzymes’ spectra beyond the usual blue range has become an attractive goal for potential industrial use of photoenzymes for chemical synthesis. However, making such predictions using quantum chemical methods can be computationally intractable, especially if the correct description of the photoexcitation requires multireference methods. We present a sequence-transferable machine learning framework, LIGand-based PROtein-agnostic SPECTral prediction (LIG-PROSPECT), to predict the absorption spectra of flavin-dependent photoenzyme variants in the presence of specific substrates. We make use of structural descriptors of cofactor-substrate ensembles as a proxy to represent the 3D conformational features of protein sequences to enable the classification of cyan-light and red-light-absorbing peaks for different conformations and the prediction of spectral features of the proteins using linear regression, offering resource-efficient alternatives to multireference quantum mechanical calculations for screening protein sequences for desired photophysical properties, validated with variants of a flavoprotein, Gluconobacter ene-reductase (GluER). With models trained using data from only two protein variants, our approach allowed us to make predictions for a protein variant that represents an extrapolation beyond the training set.","source_metadata":{"short_name":"JCTC","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"journals:8d99c4543d57572180d3e0ab964696c933171675","kind":"journals","source":"Computers and Artificial Intelligence","title":"MFECL-DTI: Multiscale Feature Extraction and Cross-Branch Contrastive Learning for Drug–Target Interaction Prediction","url":"https://doi.org/10.70267/cai.v3n6.141151","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.70267%2Fcai.v3n6.141151","date":"2026-11-18T00:00:00Z","timestamp":1794960000,"categories":[],"topic_ids":[],"keywords":["convolutional neural","Transformer"],"matched_keywords":["convolutional neural","Transformer","Drug","molecular","protein"],"matched_tags":["neural_network","transformer","generic_chem","biology"],"doi":"10.70267/cai.v3n6.141151","external_id":"8d99c4543d57572180d3e0ab964696c933171675","pdf_url":null,"code_url":null,"code_host":null,"authors":["Ye-Xin Chen","Run Zhong"],"journal":"Computers and Artificial Intelligence","publisher":null,"impact_factor":null,"abstract":"Accurate drug–target interaction prediction plays an important role in drug discovery and repositioning. However, existing methods often struggle to fully exploit the complementary information between molecular structures and sequence representations. To address this issue, we propose MFECL-DTI, a dual-branch prediction framework integrating graph-based structural encoding, multiscale sequence feature extraction, cross-branch contrastive learning, and adaptive gated fusion. Specifically, the graph branch employs residual graph convolutional networks to extract structural features from drug molecular graphs and protein contact graphs, while the sequence branch combines multiscale convolutional neural networks with a Transformer to capture local patterns and contextual dependencies. Furthermore, a bidirectional contrastive learning mechanism is introduced to align the representations of the same drug–target pair across different branches. An adaptive gating mechanism then dynamically integrates complementary features for interaction prediction. Experiments on BioSNAP, DrugBank, and C.elegans achieve AUROC values of 0.925, 0.915, and 0.994, respectively, with corresponding AUPRC values of 0.927, 0.913, and 0.994. Comparative and ablation experiments demonstrate the effectiveness of the proposed framework in improving DTI prediction performance.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:10.1038/s43588-026-01076-9","kind":"journals","source":"Nature Computational Science","title":"Minimizing data requirements for chemical and material property prediction via cluster-based training set selection","url":"https://doi.org/10.1038/s43588-026-01076-9","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1038%2Fs43588-026-01076-9","date":"2026-11-18T00:00:00+00:00","timestamp":1794960000,"categories":[],"topic_ids":[],"keywords":["property prediction"],"matched_keywords":["property prediction","chemical"],"matched_tags":["ml_task","generic_chem"],"doi":"10.1038/s43588-026-01076-9","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Quinn M. Gallagher","Sonia Hasko","Michael A. Webb"],"journal":"Nature Computational Science","publisher":"Nature","impact_factor":18.3,"abstract":"","source_metadata":{"short_name":"Nat. Comput. Sci.","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"preprints:10.26434/chemrxiv.15010421/v2","kind":"preprints","source":"ChemRxiv","title":"OPRD-target: A Collection of Manufacturing Chemistry Targets for Benchmarking Retrosynthesis Tools","url":"https://doi.org/10.26434/chemrxiv.15010421/v2","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.26434%2Fchemrxiv.15010421%2Fv2","date":"2026-11-18T00:00:00Z","timestamp":1794960000,"categories":["Cheminformatics","Reaction Informatics"],"topic_ids":["cheminformatics","reaction"],"keywords":["SMILES","multi step synthesis","synthesis planning","Retrosynthesis"],"matched_keywords":["SMILES","multi-step synthesis","Retrosynthesis Tools Modern computer-aided synthesis planning (CASP) tools, particularly retrosynthesis planning","chemical","molecule","molecules","synthesis planning","Retrosynthesis"],"matched_tags":["molecular_notation","multistep_synthesis","retrosynthesis_method","generic_chem","synthesis_planning","retrosynthesis"],"doi":"10.26434/chemrxiv.15010421/v2","external_id":"10.26434/chemrxiv.15010421/v2","pdf_url":null,"code_url":"https://github.com/KagakuLab/OPRD-target","code_host":"GitHub","authors":["Dmitry Zankov","Mario Barbatti"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"Modern computer-aided synthesis planning (CASP) tools, particularly retrosynthesis planning, are becoming practical aids for the ideation stage of chemical synthesis design. However, the same target molecule can be synthesized via substantially different routes depending on the purpose: a discovery or medicinal chemistry route used to obtain small quantities during early development, or a manufactur-ing route optimized for industrial-scale production. While CASP tools have been extensively studied and benchmarked for the former, their ability to support manufacturing route design remains largely unexplored. Performance of CASP tools is usually measured by benchmark experiments, which typi-cally report the fraction of target molecules for which a tool can generate a synthesis route, connecting the target molecule through a sequence of intermediates to starting materials. Sometimes, to assess the quality of the generated routes, they are also manually compared with expert-designed routes reported in the literature. Currently, public benchmark datasets focused on process and manufacturing chemistry are lacking. To address this gap, we take a first step toward developing a benchmark for evaluating CASP tools in manufacturing chemistry by systematically analyzing publications in Organic Process Research & Development (OPRD). We reviewed OPRD papers published between 1997 and 2022 and selected 1915 papers reporting multi-step synthesis routes. We then separately collected and digitized the target molecules of corresponding synthesis routes as SMILES strings. This resulted in a collection of 2219 unique target molecules, each paired with the DOI of the original publication reporting their corresponding synthesis route. This benchmark can therefore be used to assess how many process-chemistry targets modern CASP tools can solve and, for solved targets, to compare predicted routes with the expert-designed ones reported in the literature. The dataset is publicly available at: https://github.com/KagakuLab/OPRD-target.","source_metadata":{"date_field":"posted","keywords":[],"code_url":"https://github.com/KagakuLab/OPRD-target","code_status":"found"}},{"id":"preprints:10.26434/chemrxiv.15012450/v1","kind":"preprints","source":"ChemRxiv","title":"Pairwise learning of cyclic peptide membrane permeability on a matched-molecular-pair graph","url":"https://doi.org/10.26434/chemrxiv.15012450/v1","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.26434%2Fchemrxiv.15012450%2Fv1","date":"2026-11-18T00:00:00Z","timestamp":1794960000,"categories":["Cheminformatics"],"topic_ids":["cheminformatics"],"keywords":["matched molecular pair","matched molecular pairs","graph neural network"],"matched_keywords":["matched molecular pair","matched molecular pairs","neural network","graph neural network","Deep learning","molecular","drug","molecules"],"matched_tags":["mmp","neural_network","gnn","machine_learning","generic_chem"],"doi":"10.26434/chemrxiv.15012450/v1","external_id":"10.26434/chemrxiv.15012450/v1","pdf_url":null,"code_url":null,"code_host":null,"authors":["Wei Liu","Hwee Kuan Lee","James Lu","Jianguo Li"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"Cyclic peptides are a promising drug scaffold that can target proteins considered undruggable by small molecules, but their poor membrane permeability limits further application. Deep learning has been used to lower the wet-lab burden of finding cell-permeable peptides. Most existing models are trained on measurements curated from many publications and predict a single peptide’s permeability directly. That formulation ignores systematic measurement differences between publications, and it does not answer the actual question drug development asks: not whether a peptide is permeable, but which modification would make a given lead peptide more permeable. Here we introduce CycPeptMMP, which learns the permeability difference within a matched molecular pair (denoted as pair mode) instead of a single peptide’s value (denoted as single mode). Conventional molecular pairing approaches do not apply to cyclic peptides, so we constructed a graph of matched molecular pairs over the four modification types commonly used to improve permeability, organizing 5,674 peptides across seven publication sources. We trained both single and pair modes on five graph neural network backbones across 10 random splits. Pair mode performs better on sources with fewer than 1,000 peptides except Wang, whose peptides often lack a close analogue. Pair mode’s advantage is greatest for the closest analogues differing by a single modification. Single mode is preferable for sources with more than 1,000 peptides. We also implemented a censored loss to retain peptides with clamped measurements in the regression. For pairs involving those peptides, pair mode is more accurate when predicting which one in the pair is more permeable. Taken together, we provide a framework for comparing single and pair mode within each source, and show where pair mode is worth using.","source_metadata":{"date_field":"posted","keywords":[]}},{"id":"journals:10.1021/acs.jcim.6c03036","kind":"journals","source":"Journal of Chemical Information and Modeling","title":"PolymerFFKit: Bridging\nPolymer Chemistry and Simulation-Ready\nInputs through Traceable Force-Field and Topology Generation","url":"https://doi.org/10.1021/acs.jcim.6c03036","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1021%2Facs.jcim.6c03036","date":"2026-11-18T00:00:00+00:00","timestamp":1794960000,"categories":["Property Prediction","Targets & Structures","Spectra & Analytical"],"topic_ids":["property","targets","chemometrics"],"keywords":[],"matched_keywords":[],"matched_tags":["molecular_dynamics","generic_chem"],"doi":"10.1021/acs.jcim.6c03036","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Tong Shi","Qi Wang","Deyang Chen","Tianfu Zhang","Yingchun Liu"],"journal":"Journal of Chemical Information and Modeling","publisher":"ACS","impact_factor":5.3,"abstract":"Polymer molecular dynamics (MD) input preparation often separates experimental characterization, resolution mapping, bonded-parameter fitting, topology construction, and engine-specific file writing into weakly connected steps. PolymerFFKit is a cross-platform application that connects these steps for polymer force-field assembly and topology generation. It supports all-atom (AA), united-atom (UA), and coarse-grained (CG) representations; extracts bonded-coordinate statistics; obtains potential-of-mean-force (PMF) profiles by Boltzmann inversion; records fitting-domain choices; builds sequence-controlled polymer topologies; and exports GROMACS- and LAMMPS-oriented input files. PolymerFFKit v1.0.0 is available at https://tongshi0305-cpu.github.io/. Its main contribution is traceable preparation: mapping rules, fitted domains, random seeds, sequence definitions, and writer outputs remain part of the preparation record. In a hydroxyl-terminated polybutadiene (HTPB) case study, gel permeation chromatography (GPC), nuclear magnetic resonance (NMR), and Fourier transform infrared spectroscopy (FTIR) measurements defined sample-specific molar-mass and microstructure descriptors, while differential scanning calorimetry (DSC) provided an independent glass transition temperature (Tg). PolymerFFKit-derived coarse-grained simulations preserved the experimental Tg ranking across three independent runs for each sample. The above results demonstrate that PolymerFFKit establishes a high-quality, standardized workflow for polymer simulation, bridging experimental synthesis and characterization with the prediction of polymer properties.","source_metadata":{"short_name":"JCIM","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"journals:10.1021/acs.jcim.6c03622","kind":"journals","source":"Journal of Chemical Information and Modeling","title":"Predicting Mutation-Induced Alterations in Protein\nSolubility Using an Attention-Guided Multimodal Learning Method","url":"https://doi.org/10.1021/acs.jcim.6c03622","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1021%2Facs.jcim.6c03622","date":"2026-11-18T00:00:00+00:00","timestamp":1794960000,"categories":["Cheminformatics"],"topic_ids":["cheminformatics"],"keywords":[],"matched_keywords":[],"matched_tags":["biology"],"doi":"10.1021/acs.jcim.6c03622","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Qingxu Deng","Yao Song","Xinming Wu","Fuping Lu","Fufeng Liu"],"journal":"Journal of Chemical Information and Modeling","publisher":"ACS","impact_factor":5.3,"abstract":"Protein solubility plays a crucial role in recombinant protein production and protein developability. However, the vast mutational space makes exhaustive experimental screening costly and time-consuming, motivating computational approaches to prioritize candidate mutations. Predicting mutation-induced solubility changes remains challenging because mutation effects are associated with both the sequence context and the local three-dimensional physicochemical environment. Although multimodal models can integrate complementary sequence- and structure-related representations, the predictive contribution and biological meaning of individual architectural components cannot be inferred directly from the learned fusion behavior. Herein, we developed AttenSol, an attention-guided multimodal framework that integrates mutation-sensitive sequence representations with topology-informed graph processing through a learned fusion mechanism. On the held-out benchmark, AttenSol achieved a balanced accuracy of 52.8%. Control experiments showed that AttenSol did not exhibit a statistically significant balanced accuracy (BACC) advantage over the evaluated direct-fusion and topology-informed GCN comparators, while the learned fusion gate showed a strong but nearly constant preference for the graph-processed branch. Fixed-gate and topology-perturbation analyses indicated that this weighting should not be interpreted as evidence of structural information dominance or strong sample-adaptive routing. Forward–reverse analysis showed strong directional consistency when reverse-mutation augmentation was used, whereas matched retraining without this augmentation reversed the correlation pattern and substantially increased the zero-sum error. Cross-data set evaluation on SoluProtMutDB yielded a BACC of 0.397, revealing a substantial performance reduction under data set and domain shift. Together, these results position AttenSol as a multimodal framework for mutation effect prediction and computational prioritization while emphasizing the need for explicit control analyses when interpreting learned fusion behavior.","source_metadata":{"short_name":"JCIM","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"journals:10.1021/acs.jcim.6c03049","kind":"journals","source":"Journal of Chemical Information and Modeling","title":"Prospective Execution\nof a Generative Peptide Design\nWorkflow at the Keap1–Nrf2 Interface","url":"https://doi.org/10.1021/acs.jcim.6c03049","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1021%2Facs.jcim.6c03049","date":"2026-11-18T00:00:00+00:00","timestamp":1794960000,"categories":["Cheminformatics"],"topic_ids":["cheminformatics"],"keywords":[],"matched_keywords":[],"matched_tags":["bioactivity"],"doi":"10.1021/acs.jcim.6c03049","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Hye Ree Yoon","Kyoung Pyo Kwon","Jin Hee Lee","Min Sun Yeom"],"journal":"Journal of Chemical Information and Modeling","publisher":"ACS","impact_factor":5.3,"abstract":"Deep-learning tools for peptide design and structure prediction are increasingly combined into screening workflows, but at the Keap1–Nrf2 interface such work has rarely progressed beyond computational characterization. We report a case study in which an initial-round workflow built from publicly available components─ProteinMPNN, Chai-1, PLIP, and HADDOCK3─was fixed before any experimental data were obtained and its complete prioritized output was carried to synthesis and measurement: 530,299 unique sequences from seven Nrf2-derived backbones were reduced to 5,118 by sequence filtering and to 28 by structure-based filtering, and all 28 were synthesized and assayed alongside 16 analogues and the native Nrf2 12-mer, giving 45 measured peptides. One design, a 12-mer carrying an ENGE variant of the ETGE motif, inhibited the interaction with an apparent IC50 of 120 nM under the assay conditions used, whereas the 14 resampled analogues satisfying the same structural criteria showed low inhibition, indicating that within this workflow the structure-based steps acted as filters for structural plausibility rather than as predictors of activity. Because the set was measured in full and the intermediate sequence sets were retained, we also report how the composition of the surviving population changed across the cascade; the original workflow data are released.","source_metadata":{"short_name":"JCIM","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"journals:10.1038/s41467-026-79241-x","kind":"journals","source":"Nature Communications","title":"Quantum-accurate vibrational and conformational dynamics from small molecules to proteins with machine-learned force fields","url":"https://doi.org/10.1038/s41467-026-79241-x","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1038%2Fs41467-026-79241-x","date":"2026-11-18T00:00:00+00:00","timestamp":1794960000,"categories":[],"topic_ids":[],"keywords":["machine learned force fields","DFT","force fields"],"matched_keywords":["machine-learned force fields","DFT","force fields","molecules","molecular","protein"],"matched_tags":["ml_force_field","dft","force_fields","generic_chem","biology"],"doi":"10.1038/s41467-026-79241-x","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Sergio Suárez-Dou","Miguel Gallegos","Kyunghoon Han","Florian N. Brünig","Joshua T. Berryman","Alexandre Tkatchenko"],"journal":"Nature Communications","publisher":"Nature","impact_factor":15.7,"abstract":"Biomolecular thermodynamics and spectroscopy depend on relative conformer energies, local curvatures, and collective dipole fluctuations on the potential-energy surface. Conventional molecular mechanics force fields enable large-scale simulations, but their fixed functional forms can misrepresent infrared intensities, mode character, and environment-dependent vibrational response. Here we assess general-purpose machine-learned force fields by using finite-temperature infrared spectra across small molecules, gas-phase peptides, and monomeric, oligomeric, and solvated protein assemblies. To enable this analysis, we introduce QVib, a dataset of 293 molecules and 1366 conformers, together with peptide amide-band benchmarks and p53 oligomerization-domain models, to evaluate vibrational transferability from DFT references to experimental spectra. Across these systems, machine-learned force fields improve over molecular mechanics in reproducing DFT-level forces, vibrational frequencies, densities of states, mode eigenvectors, conformational energetics, and experimental infrared spectra. Among models with explicit long-range electrostatics, SO3LR provides the best accuracy–cost balance of the models tested here. These results show that machine-learned force-field dynamics can recover collective, environment-dependent vibrational landscapes at near-DFT fidelity, enabling spectroscopically validated biomolecular simulations at force-field-like cost.","source_metadata":{"short_name":"Nat. Commun.","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"journals:f422c6013a280047751bfdb2289ca72536540f1a","kind":"journals","source":"Oncologie","title":"Research progress on Raman spectroscopy-based noninvasive detection of tumor-derived exosomes in urine, saliva, and sweat","url":"https://doi.org/10.1515/oncologie-2026-0138","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1515%2Foncologie-2026-0138","date":"2026-11-18T00:00:00Z","timestamp":1794960000,"categories":["Cheminformatics"],"topic_ids":["cheminformatics"],"keywords":["chemometric","molecular fingerprints"],"matched_keywords":["chemometric","molecular fingerprints","molecular"],"matched_tags":["chemometrics","molecular_fingerprints","generic_chem"],"doi":"10.1515/oncologie-2026-0138","external_id":"f422c6013a280047751bfdb2289ca72536540f1a","pdf_url":null,"code_url":null,"code_host":null,"authors":["Jia-Yang Wang","Bo-Yu Zhao","Bing-Zhou Jiang","Shao-Shuai Han","Chang-Gui Tong","Zhen-Lin Wu"],"journal":"Oncologie","publisher":null,"impact_factor":null,"abstract":"Early cancer diagnosis urgently requires reproducible and minimally invasive approaches for dynamic monitoring. Exosomes are stably present in biological fluids, and they are relatively abundant, making them promising biomarkers for liquid biopsy. Raman spectroscopy, particularly surface-enhanced Raman scattering (SERS), enables the acquisition of global molecular fingerprints of exosomes within complex biofluid backgrounds, thereby providing a feasible strategy for noninvasive detection in urine, saliva, and sweat. Studies conducted under controlled laboratory conditions have demonstrated discriminatory potential. However, substantial challenges remain for clinical translation, including insufficient reproducibility, limited comparability due to variations in biofluids and substrates, inconsistencies in instrumentation and analytical workflows, and the lack of standardized modeling strategies. These issues hinder the extrapolation of high-performance experimental results to real-world clinical settings. In this review, key advances in exosome isolation, SERS substrates, signal acquisition, and chemometric or machine-learning-based data analysis are systematically summarized. The applicability boundaries of these approaches across different biofluids and clinical scenarios are evaluated. The biological interpretability of global spectral fingerprints and the reliability of small-sample modeling are discussed. Finally, potential optimization strategies, including standardized workflows, multimodal data integration, and multicenter validation, are proposed to facilitate the clinical translation of noninvasive SERS-based exosome detection.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:10.1021/jacs.6c14850","kind":"journals","source":"Journal of the American Chemical Society","title":"Revealing Bulk-Phase\nReactivity of Fe-Based Catalyst\nin the Reverse Water–Gas Shift Reaction","url":"https://doi.org/10.1021/jacs.6c14850","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1021%2Fjacs.6c14850","date":"2026-11-18T00:00:00+00:00","timestamp":1794960000,"categories":[],"topic_ids":[],"keywords":["MLIP","reaction conditions","molecular dynamics"],"matched_keywords":["MLIP","molecular","reaction conditions","molecular dynamics"],"matched_tags":["mlip_tools","generic_chem","reaction_conditions","molecular_dynamics"],"doi":"10.1021/jacs.6c14850","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Haonan Yu","Ze-Yu Li","Hanzi Liu","Wei-Wei Wang","Zhiping Zhu","Zhiqiang Sun","Chun-Jiang Jia","Jin-Cheng Liu"],"journal":"Journal of the American Chemical Society","publisher":"ACS","impact_factor":15.7,"abstract":"Heterogeneous catalysis is conventionally interpreted as a surface-localized process, yet many catalysts undergo extensive reaction-driven restructuring that may fundamentally redistribute reactivity beyond the interface. Here we combine microscopy, isotope-labeling experiments, and large-scale molecular dynamics driven by a machine-learning interatomic potential (MLIP-MD) to investigate reactive carbon–oxygen chemistry in iron catalysts relevant to the reverse water–gas shift reaction. We show that under reaction conditions, the catalyst evolves into a dynamic FeOx/Fe3C mixed phase in which carbon and oxygen continuously redistribute across surface, subsurface and bulk regions. Isotope experiments reveal that bulk carbon is catalytically active, while simulations identify initial C–O association across a broad depth range, including within the catalyst interior, for carbon atoms subsequently removed as CO/CO2. Free-energy calculations characterize representative C–O association pathways at the surface and within the catalyst interior. Together, these results support CO formation beyond conventional surface sites and a surface-to-bulk catalytic picture for dynamically evolving iron catalysts.","source_metadata":{"short_name":"JACS","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"journals:10.1021/acs.jcim.6c02123","kind":"journals","source":"Journal of Chemical Information and Modeling","title":"ScentOR: Semantic-Only\nHeterogeneous Knowledge Distillation\nfor Odorant Receptor Deorphanization","url":"https://doi.org/10.1021/acs.jcim.6c02123","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1021%2Facs.jcim.6c02123","date":"2026-11-18T00:00:00+00:00","timestamp":1794960000,"categories":["Cheminformatics"],"topic_ids":["cheminformatics"],"keywords":["gradient boosting","Receptor"],"matched_keywords":["neural network","gradient boosting","ligands","Chemical","ligand","Receptor","protein"],"matched_tags":["neural_network","ensemble","generic_chem","biology"],"doi":"10.1021/acs.jcim.6c02123","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Eun Cheol Kim","YounJoon Jung"],"journal":"Journal of Chemical Information and Modeling","publisher":"ACS","impact_factor":5.3,"abstract":"Predicting interactions between odorant ligands and odorant receptor (OR) proteins is hindered by the scarcity of experimental structures and the wide sequence divergence within the OR family. We introduce ScentOR (Semantic Chemical Encoder via Neural Transfer for Odorant Receptors), a heterogeneous knowledge distillation framework that transfers the predictive distributions of a gradient boosting (GB) teacher into a neural network (NN) student using only one-dimensional (1D) semantic representations, achieving an area under the receiver operating characteristic curve (AUROC) of 0.954 and a Matthews correlation coefficient (MCC) of 0.602 on the curated M2OR data set of 39,723 odorant ligand-OR protein pairs, matching the current best-performing model while reducing the run-to-run standard deviation approximately 3-fold. A systematic comparison of eight scenarios shows that statically predicted three-dimensional (3D) topologies, encoded as geometry alone and at the accuracy currently attainable for ORs, add no complementary value: student models assign them near-zero importance, and the small gain they produce without distillation does not survive it. We further characterize what post hoc attribution reports for a model of this design, showing that integrated gradients (IG) and in silico alanine mutagenesis concentrate on the family-conserved DRY motif arginine yet carry no pair-specific information, even though pair-specific interaction dominates the predictions.","source_metadata":{"short_name":"JCIM","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"preprints:10.26434/chemrxiv.15012433/v1","kind":"preprints","source":"ChemRxiv","title":"Subspace Fokker–Planck Transformer: An Ensemble-Based Framework for Accelerated Free-Energy Sampling","url":"https://doi.org/10.26434/chemrxiv.15012433/v1","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.26434%2Fchemrxiv.15012433%2Fv1","date":"2026-11-18T00:00:00Z","timestamp":1794960000,"categories":[],"topic_ids":[],"keywords":["Transformer"],"matched_keywords":["ML","Transformer","molecular"],"matched_tags":["machine_learning","transformer","generic_chem"],"doi":"10.26434/chemrxiv.15012433/v1","external_id":"10.26434/chemrxiv.15012433/v1","pdf_url":null,"code_url":null,"code_host":null,"authors":["Zichen Sun","Xiaoxuan Yu","Yi Qin Gao","Jun Zhang"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"Rare transitions between metastable states limit the convergence of molecular free-energy calculations. Here, we introduce the subspace Fokker–Planck Transformer (sFPT), an ensemble-based enhancedsampling method that transports the instantaneous replica distribution in a selected collective-variable (CV) subspace. sFPT estimates the score of the evolving CV distribution using a kernel representation of the replica ensemble, maps the resulting entropy force to atomic coordinates through the CV Jacobian, and incorporates this force into underdamped Langevin dynamics. We show that the self-consistent stationary CV distribution satisfies PCV ∞ (s) ∝ [PCV phys(s)] λ/(1+λ), yielding a controllably flattened free-energy surface, and establish convergence of the ideal mean-field dynamics through a nonincreasing Lyapunov functional. Post-convergence trajectories can be accumulated to improve the statistical precision of freeenergy reconstruction. We evaluate sFPT on alanine dipeptide in vacuum and explicit water, alanine tetrapeptide in a three-dimensional CV space, and a Claisen rearrangement described by an ML/MM potential in explicit water. The reconstructed free-energy landscapes agree with umbrella-sampling, replica-exchange, and well-tempered metadynamics (WT-MetaD) references. For alanine dipeptide in vacuum and explicit water, comparisons at matched walker numbers show that both methods can benefit from parallel sampling. For the Claisen rearrangement, sFPT reaches sustained agreement with the reference basin free-energy difference after approximately 24 ps per replica, whereas the selected 64-walker WT-MetaD calculation retains slow deviations over 400 ps per walker. This advantage remains visible when the integration steps over all replicas are counted. These results establish instantaneous ensemble transport as a practical route to stable free-energy estimates from many short parallel trajectories.","source_metadata":{"date_field":"posted","keywords":[]}},{"id":"journals:10.1038/s41467-026-78403-1","kind":"journals","source":"Nature Communications","title":"Supervised deep learning for efficient cryo-EM image alignment in drug discovery with cryoPARES","url":"https://doi.org/10.1038/s41467-026-78403-1","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1038%2Fs41467-026-78403-1","date":"2026-11-18T00:00:00+00:00","timestamp":1794960000,"categories":[],"topic_ids":[],"keywords":[],"matched_keywords":["deep learning","drug"],"matched_tags":["machine_learning","generic_chem"],"doi":"10.1038/s41467-026-78403-1","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Ruben Sanchez-Garcia","Alex Berndt","Amir Apelbaum","Judith Reeks","Pamela A. Williams","Carl Poelking","Charlotte M. Deane","Michael Saur"],"journal":"Nature Communications","publisher":"Nature","impact_factor":15.7,"abstract":"","source_metadata":{"short_name":"Nat. Commun.","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"journals:10.1021/acsomega.6c10589","kind":"journals","source":"ACS Omega","title":"Symmetry and Anion\nComposition Effects on HDABCO-based\nFerroelectric Ionic Plastic Crystals","url":"https://doi.org/10.1021/acsomega.6c10589","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1021%2Facsomega.6c10589","date":"2026-11-18T00:00:00+00:00","timestamp":1794960000,"categories":[],"topic_ids":[],"keywords":["machine learned force fields","density functional theory","DFT","force fields","molecular dynamics"],"matched_keywords":["machine-learned force fields","density functional theory","DFT","force fields","molecular","compounds","ligand","molecular dynamics"],"matched_tags":["ml_force_field","dft","force_fields","generic_chem","molecular_dynamics"],"doi":"10.1021/acsomega.6c10589","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Sindre Mjøen Svendsen","Didrik R. Småbråten","Mari-Ann Einarsrud","Sverre M. Selbach","Julian Walker"],"journal":"ACS Omega","publisher":"ACS","impact_factor":null,"abstract":"Ferroelectric ionic plastic crystals represent an exciting opportunity for nontoxic electrical energy storage materials. While plastic crystals exhibit a high-temperature mesophase in which the molecular components have rotational freedom, our understanding and control of molecular rotation are limited. Here, we have systematically changed the anion in the ferroelectric ionic plastic crystal [HDABCO]ClO4, with smaller and larger anionic species, BF4– and ReO4–, respectively. Using machine-learned force fields (MLFFs) derived from density functional theory (DFT), we have performed molecular dynamics (MD) simulations on the compounds, in order to investigate anion rotation and proton disorder on the atomistic scale. The rotation of the anions goes through three distinct phases; oscillation around minima positions, ligand reorientation motion, and full rotation in the high-temperature phase, where the rotational speed is correlated with anion size and crystal packing fraction. The proton transfer is calculated to reach THz frequency range at room temperature due to the low energy barrier for proton hopping, and both [HDABCO]ClO4 and [HDABCO]BF4 show significant positional disorder of the proton below the phase transition temperature. Anion substitution can enable tuning of the rotational and proton positional order of HDABCO-based ionic plastic crystals.","source_metadata":{"short_name":"ACS Omega","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"journals:10.1021/acs.jcim.6c02916","kind":"journals","source":"Journal of Chemical Information and Modeling","title":"TH-CADD: An Integrated\nHigh-Performance Computing\nPlatform for Large-Scale Computer-Aided Drug Discovery","url":"https://doi.org/10.1021/acs.jcim.6c02916","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1021%2Facs.jcim.6c02916","date":"2026-11-18T00:00:00+00:00","timestamp":1794960000,"categories":["Docking & Screening"],"topic_ids":["docking"],"keywords":["molecular docking","molecular dynamics"],"matched_keywords":["molecular docking","Drug","chemical","ligand","molecular","molecular dynamics","protein"],"matched_tags":["docking","generic_chem","molecular_dynamics","biology"],"doi":"10.1021/acs.jcim.6c02916","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Peishun Li","Feifei Xiao","Qi Wu","Pengfei Nie","Liangsheng Chen","Xianze Zhang","Xiaodong Jian","Boyang Ji","Youtao Song","Xiangfei Meng"],"journal":"Journal of Chemical Information and Modeling","publisher":"ACS","impact_factor":5.3,"abstract":"The growing scale of chemical libraries and the increasing complexity of biological targets have made high-throughput computer-aided drug design (CADD) an essential component of modern drug discovery. Here, we present TH-CADD, a high-performance computer-aided drug discovery platform built on the New Generation Tianhe supercomputer. TH-CADD integrates target preparation and analysis, ligand preparation, similarity search, molecular docking, postdocking analysis, and molecular dynamics simulation into a unified computational framework. A Django-Vue architecture deployed on Tianhe Cloud provides web-based access to high-performance computing (HPC) resources through scheduler-controlled execution. By organizing established CADD tools into a coherent workflow, TH-CADD supports accessible, traceable, and scalable computational drug discovery. The platform supports multiple molecular input formats, automated preprocessing, large-scale screening, and molecular dynamics evaluation of selected protein–ligand complexes. TH-CADD’s contribution lies in the system-level integration, automated job management, and scalable execution of established computational methods on the New Generation Tianhe supercomputer, together with a unified workflow view that supports direct cross-module reuse of compatible task outputs.","source_metadata":{"short_name":"JCIM","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"preprints:10.26434/chemrxiv.15012417/v1","kind":"preprints","source":"ChemRxiv","title":"Understanding a Helix Reverse Path in Perfluoroalkanes from a Neural Network Potential","url":"https://doi.org/10.26434/chemrxiv.15012417/v1","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.26434%2Fchemrxiv.15012417%2Fv1","date":"2026-11-18T00:00:00Z","timestamp":1794960000,"categories":[],"topic_ids":[],"keywords":["Neural Network Potential","MACE","DFT"],"matched_keywords":["Neural Network Potential","MACE","Neural Network","DFT","molecule","chemical"],"matched_tags":["mlip_tools","neural_network","dft","generic_chem"],"doi":"10.26434/chemrxiv.15012417/v1","external_id":"10.26434/chemrxiv.15012417/v1","pdf_url":null,"code_url":null,"code_host":null,"authors":["Shun Konno","Takeshi Hasewgawa"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"The helical handedness of perfluoroalkanes and polytetrafluoroethylene (PTFE) is considered to reverse through localized helix-reversal defects. Earlier studies with empirical and semiempirical potentials described how such defects form and move. The helix, however, originates from hyperconjugation, and the reversal pathway of a single molecule has not yet been examined with a potential that includes this quantum-chemical effect. Here we trained a MACE potential on DF-MP2-F12 data for C4F10–C7F16 and located three reversal pathways of C24F50 by CI-NEB: a concerted pathway through the all-trans (D2h) structure, a single-wall pathway, and a twin-wall pathway. For C24F50, the concerted barrier is 13.74 kcalmol−1, far above those of the single-wall pathway (2.01 kcalmol−1) and the twin-wall pathway (4.82 kcalmol−1); the concerted barrier grows linearly with chain length, whereas the single-wall barrier does not depend on it. At 25 ◦C, the single-wall pathway is about 108 times faster than the concerted one for C24F50 when the prefactors are assumed equal, so the concerted reversal of an isolated chain is kinetically very unfavorable. For C13F28–C15F32, DFT calculations show that the frequency of the wall mode depends on the position of the wall, so wall-containing conformations can contribute to the broad defectrelated band in the ATR-IR spectra of these short-chain perfluoroalkanes. The growth of the low-wavenumber component from 20 to 70 ◦C is consistent with an increasing thermal population of these conformations. In the single-wall path, a moving wall also rotates the chain about its axis. Combined with crystal simulations in the literature, this supports the screw-like motion proposed for PTFE, in which the rotation of a chain is accompanied by a translation along its axis.","source_metadata":{"date_field":"posted","keywords":[]}},{"id":"preprints:10.26434/chemrxiv.15012453/v1","kind":"preprints","source":"ChemRxiv","title":"Using potential derivatives to reduce the cost of rectangular collocation: a general and versatile approach for computational vibrational spectroscopy without a potential energy surface","url":"https://doi.org/10.26434/chemrxiv.15012453/v1","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.26434%2Fchemrxiv.15012453%2Fv1","date":"2026-11-18T00:00:00Z","timestamp":1794960000,"categories":[],"topic_ids":[],"keywords":["Claude","DFT","force field","ab initio"],"matched_keywords":["Claude","DFT","force field","ab initio","molecule"],"matched_tags":["LLMs","dft","force_fields","quantum_chemistry","generic_chem"],"doi":"10.26434/chemrxiv.15012453/v1","external_id":"10.26434/chemrxiv.15012453/v1","pdf_url":null,"code_url":null,"code_host":null,"authors":["Moksh Bhateja","Tucker Carrington","Sergei Manzhos"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"We modify the exact Cartesian kinetic energy operator, rectangular collocation approach of Manzhos and Carrington [J. Chem. Phys. 145 (2016) 224110] by using information contained in potential derivatives to decrease the number of required potential points. We show that this makes it possible to substantially reduce the number of collocation points, which is important, in particular, when energies are computed directly ab initio. Vibrational band intensities are also computed. We first demonstrate the method on Cl2CO and compare to high-accuracy Smolyak collocation calculations on an accurate potential energy surface (PES). Sub-cm-1 accuracy is obtained with a few thousand Gaussian basis functions and potential points. A spectrum computed directly from DFT energies and a comparison to GVPT2 (generalized vibrational perturbation theory) are also presented. We then consider LiCO2 as a more anharmonic case, for which a quartic force field potential is less reliable, and which also serves as a surrogate model for a catalytic system (mimicking CO2 bond weakening by charge donation). The rectangular collocation method is used to compute the spectrum of LiCO2 directly from ab initio energies and to compare to GVPT2. The method is implemented as a Claude Skill and is easily usable by anyone on any molecule. The method is a general and versatile tool for anharmonic computational vibrational spectroscopy. It does not require a PES and no coupling is neglected.","source_metadata":{"date_field":"posted","keywords":[]}},{"id":"journals:10.1021/acs.jmedchem.6c01918","kind":"journals","source":"Journal of Medicinal Chemistry","title":"What Does Generalizable AI in Drug Discovery Really Mean?","url":"https://doi.org/10.1021/acs.jmedchem.6c01918","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1021%2Facs.jmedchem.6c01918","date":"2026-11-18T00:00:00+00:00","timestamp":1794960000,"categories":[],"topic_ids":[],"keywords":[],"matched_keywords":["Machine learning","Drug"],"matched_tags":["machine_learning","generic_chem"],"doi":"10.1021/acs.jmedchem.6c01918","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Benjamin P. Brown"],"journal":"Journal of Medicinal Chemistry","publisher":"ACS","impact_factor":6.8,"abstract":"Machine learning models increasingly contribute to drug discovery, but their utility depends on performing reliably in their intended deployment setting. Although model generalization is invoked frequently, the drug discovery community lacks a shared vocabulary for how models actually achieve transferability beyond their training data. We argue that achieving generalization requires examining how a model’s inductive bias interacts with its training data to shape the notion of similarity it learns. Without a bias that makes relevant recurring relationships easier to learn as opposed to incidental correlations, models exploit dataset-specific shortcuts. These shortcuts yield high benchmark performance but fail in prospective applications. We therefore propose evaluation principles tailored to a model’s architecture, using diagnostic controls and train−test splits that distinguish transferable learning from the retrieval of familiar training data. This Perspective shows how representational choices shape failure modes across pose-dependent, pose-independent, and foundation model paradigms, and provides a practical framework to help researchers expose shortcuts and interpret predictions.","source_metadata":{"short_name":"J. Med. Chem.","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"preprints:2611.22735v1","kind":"preprints","source":"arXiv","title":"Spin-Adapted Neural-Network Backflow Wavefunctions for Excited States of Strongly Correlated Open-Shell Systems","url":"https://arxiv.org/abs/2611.22735v1","detail_url":"/chemradar/article?u=https%3A%2F%2Farxiv.org%2Fabs%2F2611.22735v1","date":"2026-11-17T13:06:47Z","timestamp":1794920807,"categories":[],"topic_ids":[],"keywords":[],"matched_keywords":["neural network","molecular","chemical"],"matched_tags":["neural_network","generic_chem"],"doi":null,"external_id":"2611.22735v1","pdf_url":"https://arxiv.org/pdf/2611.22735v1","code_url":null,"code_host":null,"authors":["Bohan Zhang","Zibo Wu","Yunzhi Li","Wei-Hai Fang","Zhendong Li"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"Building on the success of neural-network quantum states (NQS) for molecular ground states, multi-state frameworks such as natural excited states (NES) [Pfau et al., Science 2024, 385, eadn0137] extend NQS to chemically important excited states. However, without control over total spin, such calculations can require the inclusion of many intervening states of other spins, particularly in strongly correlated systems with dense low-energy spectra. We therefore generalize our previously developed spin-adapted neural network backflow (SA-NNBF) ansatz [Li et al., arXiv:2604.06841 (2026)] to improve its variational flexibility for excited-state calculations within a prescribed spin sector. We establish the universality of this extended ansatz within the chosen orbital space and spin sector, and use numerical benchmarks to assess how parameter allocation affects the ansatz's accuracy. With the spin-adapted ansatz, we are able to directly optimize the first excited state (a singlet) of $O_2$ without using any multi-state frameworks. Combining the extended ansatz with NES, we obtain low-lying triplet states of $O_2$ and singlet states of the $[\\mathrm{Fe}_2\\mathrm{S}_2(\\mathrm{SCH}_3)_4]^{2-}$ cluster without including intervening states of other spins. Both the single-state and NES calculations achieve chemical accuracy relative to the exact energies. Our work provides a practical NQS framework for targeting excited states of prescribed spin in strongly correlated systems.","source_metadata":{"categories":["physics.chem-ph"],"code_status":"absent"}},{"id":"preprints:2611.22623v1","kind":"preprints","source":"arXiv","title":"What Do LLMs Use to Judge Synthesis Recipes? An Experimental Analysis on 240 Solid-State Syntheses","url":"https://arxiv.org/abs/2611.22623v1","detail_url":"/chemradar/article?u=https%3A%2F%2Farxiv.org%2Fabs%2F2611.22623v1","date":"2026-11-17T11:37:42Z","timestamp":1794915462,"categories":["LLMs & Agents"],"topic_ids":["llm"],"keywords":["LLMs","LLM","GPT"],"matched_keywords":["LLMs","LLM","GPT","chemical"],"matched_tags":["LLMs","transformer","generic_chem"],"doi":null,"external_id":"2611.22623v1","pdf_url":"https://arxiv.org/pdf/2611.22623v1","code_url":null,"code_host":null,"authors":["Derek van Tilborg","Laura-Bianca Pasca","Harry Sevi","Kyle Reeves","Thomas Loux","Aikaterini Flessa Savvidou","Varincka Jean-Gilles","Marilou Carouge","Joonatan E. M. Laulainen"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"Inorganic materials discovery is increasingly limited by synthesis rather than by material design. Large language models (LLMs) are being adopted to propose synthesis conditions with little experimental evidence of their capabilities and biases. Here, we carry out an experimental synthesis benchmark of 240 experiments across eight known and unknown inorganic oxides and subsequently evaluate an LLM as a zero-shot ranker over fixed banks of solid-state synthesis recipes. For each pair of recipes for the same oxide target, `GPT-5.6-Sol` selects the recipe expected to yield higher target-phase purity. We aggregate these pairwise preferences using a regularized Bradley-Terry model to obtain a ranked list of synthesis recipes for a given target material. By comparison with the actual purity, obtained through X-ray diffraction, we measure `GPT-5.6-Sol`'s performance as a zero-shot ranker (0.72 pairwise accuracy). By selective removal of the information exposed to the model during this ranking, we conclude that ranking is driven predominantly by thermal features, with reaction chemistry modulating the effect. Ranking accuracy tracks the purity-temperature correlation by target (Spearman ρ= 0.91\\), \\(p = 0.002\\), and on the six thermally-driven targets withholding all chemical identity hardly impacts performance (0.68) compared to withholding temperature (0.56). Ranking performance was higher on known than unknown targets, but the group-level difference was not statistically conclusive. The methodological choices highlight the need for more experimentally validated synthesis benchmarks and position LLMs as a thermally informed prior.","source_metadata":{"categories":["cs.CE","cond-mat.mtrl-sci"],"code_status":"absent"}},{"id":"journals:10.1039/d6dd00705h","kind":"journals","source":"Digital Discovery","title":"NEAT-POCKET: Pocket-Conditioned Autoregressive 3D Molecular Generation with a Neighborhood-Guided Set Transformer","url":"https://doi.org/10.1039/d6dd00705h","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1039%2Fd6dd00705h","date":"2026-11-17T10:43:53Z","timestamp":1794912233,"categories":["Design de novo"],"topic_ids":["denovo"],"keywords":["Molecular Generation","de novo molecular design","Transformer"],"matched_keywords":["Molecular Generation","de novo molecular design","Transformer","Molecular","drug","ligands","protein"],"matched_tags":["de_novo_design","transformer","generic_chem","biology"],"doi":"10.1039/d6dd00705h","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Roxane Axel Jacob","Daniel Rose","Thierry Langer","Johannes Kirchmair"],"journal":"Digital Discovery","publisher":"RSC","impact_factor":5.6,"abstract":"AI-driven de novo molecular design offers a promising route to accelerate early-stage drug discovery by generating novel ligands directly within target protein binding pockets. We present NEAT-POCKET, a pocket-conditioned extension...","source_metadata":{"short_name":"Digital Discovery","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"preprints:2611.23342v1","kind":"preprints","source":"arXiv","title":"Confidence-Based Materials Data Quality Assessment via Domain Knowledge and Graph-Structured Learning","url":"https://arxiv.org/abs/2611.23342v1","detail_url":"/chemradar/article?u=https%3A%2F%2Farxiv.org%2Fabs%2F2611.23342v1","date":"2026-11-17T08:06:38Z","timestamp":1794902798,"categories":["ML Potentials"],"topic_ids":["potentials"],"keywords":["MACE"],"matched_keywords":["MACE","machine learning"],"matched_tags":["mlip_tools","machine_learning"],"doi":null,"external_id":"2611.23342v1","pdf_url":"https://arxiv.org/pdf/2611.23342v1","code_url":"https://github.com/materialsCnicCas/materials-data-quality","code_host":"GitHub","authors":["Sihan Liaing","Ji Zhang","Kehan Wang","Zihao Gao","Wei Ma","Yangang Wang","Zongguo Wang"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"The rapid development of materials machine learning relies on large-scale, heterogeneous datasets, yet model reliability is often compromised by imperfect data from diverse experiments, simulations, and database integration. Real-world materials databases may contain measurement errors, failed calculations, missing or duplicate records, and hidden violations of physicochemical constraints. Such problems are difficult to identify because corrupted samples may remain statistically plausible rather than appearing as explicit outliers. In this paper, we propose a confidence-based materials data quality assessment framework that integrates domain knowledge with graph-structured learning. Data quality is formulated as a continuous confidence estimation problem rather than a binary anomaly identification task, enabling sample-level assessment and historical-reference-based batch comparison. A nine-dimensional feature-quality prior and duplicate-aware sample initialization are calibrated using materials domain rules and statistical evidence. Confidence is further refined through sample-feature bipartite graph propagation and weakly supervised graph attention, while a Memory-GAT module provides historical reference scoring. Experiments on 23 ODDS benchmark datasets show that the generic module achieves an average AUC-ROC of 0.866. In five repeated mixed-contamination experiments on mp\\_20, rule calibration improves HistGradientBoostingRegressor test $R^2$ from $0.538\\pm0.008$ to $0.616\\pm0.014$. Graph refinement increases average precision from $0.943\\pm0.004$ to 0.953\\pm0.004$, while the rule-calibrated configuration has the highest mean test $R^2$ among the evaluated screening configurations. Five repeated MACE subset experiments further show an increase in $R^2$ from $0.590\\pm0.053$ to $0.880\\pm0.027$ after rule-based screening.","source_metadata":{"categories":["cs.LG","cond-mat.mtrl-sci","cs.AI"],"code_status":"found","code_url":"https://github.com/materialsCnicCas/materials-data-quality"}},{"id":"preprints:2611.22272v1","kind":"preprints","source":"arXiv","title":"Neural Network Path Optimization for Finding Transition States on a Machine Learned Potential","url":"https://arxiv.org/abs/2611.22272v1","detail_url":"/chemradar/article?u=https%3A%2F%2Farxiv.org%2Fabs%2F2611.22272v1","date":"2026-11-17T06:18:18Z","timestamp":1794896298,"categories":["ML Potentials"],"topic_ids":["potentials"],"keywords":["MLIP","density functional theory"],"matched_keywords":["MLIP","Neural Network","density functional theory","chemical"],"matched_tags":["mlip_tools","neural_network","dft","generic_chem"],"doi":null,"external_id":"2611.22272v1","pdf_url":"https://arxiv.org/pdf/2611.22272v1","code_url":"https://github.com/deepprinciple/ReactBench","code_host":"GitHub","authors":["Kareem Hegazy","Eric C-Y Yuan","Anup Kumar","Dmitriy Morozov","Daniel Barter","Michael W Mahoney","Teresa Head-Gordon","Samuel M Blau"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"Transition states govern chemical reaction rates and outcomes, yet locating them remains a central challenge in computational chemistry. Double-ended search methods find transition states given reactant and product structures, but generally rely on discretized path representations developed for density functional theory, where potential energy surface evaluations are expensive. Machine learned interatomic potentials (MLIPs) dramatically reduce this cost and leverage automatic differentiation to access higher-order derivatives efficiently. We introduce Popcornn, a framework to find transition states by optimizing continuous neural network paths atop an MLIP with a parallelized adaptive numerical integrator. This combination eliminates path discretization artifacts, reduces hyperparameter sensitivity, and enables non-local optimization to overcome local features using second-order potential-energy-surface information. Popcornn consistently recovers more correct transition states than standard methods such as nudged elastic band and growing string, across MLIPs on thousands of organic reactions, establishing it as a robust, automation-ready method for high-throughput transition state search. Popcornn's infrastructure is flexible in path objective, integration strategy, and optimization algorithm, making it a practical tool and platform for investigating open questions in MLIP-based path optimization.","source_metadata":{"categories":["physics.chem-ph","cs.AI"],"code_status":"found","code_url":"https://github.com/deepprinciple/ReactBench"}},{"id":"preprints:2611.22230v1","kind":"preprints","source":"arXiv","title":"Discovery of new antiviral agents through artificial intelligence: in vitro and in vivo results","url":"https://arxiv.org/abs/2611.22230v1","detail_url":"/chemradar/article?u=https%3A%2F%2Farxiv.org%2Fabs%2F2611.22230v1","date":"2026-11-17T05:22:24Z","timestamp":1794892944,"categories":["Design de novo"],"topic_ids":["denovo"],"keywords":["reinforcement learning","RL","Molecular dynamics","EC50"],"matched_keywords":["reinforcement learning","RL","molecule","compounds","Molecular","Molecular dynamics","EC50"],"matched_tags":["reinforcement_learning","generic_chem","molecular_dynamics","bioactivity"],"doi":"10.1016/j.antiviral.2024.105818","external_id":"2611.22230v1","pdf_url":"https://arxiv.org/pdf/2611.22230v1","code_url":null,"code_host":null,"authors":["Roza Izmailyan","Mher Matevosyan","Hamlet Khachatryan","Anastasiya Shavina","Smbat Gevorgyan","Artur Ghazaryan","Irina Tirosyan","Yeva Gabrielyan","Marusya Ayvazyan","Boris Martirosyan","Vardan Harutyunyan","Hovakim Zakaryan"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"In this research, we employed a deep reinforcement learning (RL)-based molecule design platform to generate a diverse set of compounds targeting the neuraminidase (NA) of influenza A and B viruses. A total of 60291 compounds were generated, of which 86.5% displayed superior physicochemical properties compared to oseltamivir. After narrowing down the selection through computational filters, nine compounds with non-sialic acid-like structures were selected for in vitro experiments. We identified two compounds, DS-22-inf-009 and DS-22-inf-021 that effectively inhibited the NAs of both influenza A and B viruses (IAV and IBV), including H275Y mutant strains at low micromolar concentrations. Molecular dynamics simulations revealed a similar pattern of interaction with amino acid residues as oseltamivir. In cell-based assays, DS-22-inf-009 and DS-22-inf-021 inhibited IAV and IBV in a dose-dependent manner with EC50 values ranging from 0.29 uM to 2.31 uM. Furthermore, animal experiments showed that both DS-22-inf-009 and DS-22-inf-021 exerted antiviral activity in mice, conferring 65% and 85% protection from IAV (H1N1 pdm09), and 65% and 100% protection from IBV (Yamagata lineage), respectively. Thus, these findings demonstrate the potential of RL to generate compounds with promising antiviral properties.","source_metadata":{"categories":["q-bio.BM","cs.LG"],"code_status":"absent"}},{"id":"preprints:2611.22140v1","kind":"preprints","source":"arXiv","title":"Synthetic Crystal Data Generation for Diffusion Pretraining in Material Property Prediction","url":"https://arxiv.org/abs/2611.22140v1","detail_url":"/chemradar/article?u=https%3A%2F%2Farxiv.org%2Fabs%2F2611.22140v1","date":"2026-11-17T03:27:08Z","timestamp":1794886028,"categories":["Property Prediction","Design de novo","ML Potentials"],"topic_ids":["property","denovo","potentials"],"keywords":["graph neural networks","Property Prediction","denoising","DFT"],"matched_keywords":["neural networks","graph neural networks","Property Prediction","denoising","DFT","compounds"],"matched_tags":["neural_network","gnn","ml_task","diffusion","dft","generic_chem"],"doi":null,"external_id":"2611.22140v1","pdf_url":"https://arxiv.org/pdf/2611.22140v1","code_url":null,"code_host":null,"authors":["Aamod Varma","Shuyi Jia","Victor Fung"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"Diffusion-based pretraining has emerged as a powerful strategy for learning transferable representations in graph neural networks for materials property prediction without requiring quantum-mechanical labels such as energies, forces, and stresses for every pretraining structure. However, the method retains a critical dependency: it requires structures residing near local minima of a potential energy surface, which have traditionally been obtained through costly DFT relaxations of known or physically synthesized compounds. This reliance constrains both the scale and the structural diversity of available pretraining data, leaving large regions of crystal structure space unexplored. In this work, we connect diffusion-based pretraining with synthetic structure generation to reduce this bottleneck. We introduce a synthetic data generation pipeline combining three complementary strategies: perturb-and-relax, generative sampling via MatterGen, and protostructure enumeration. Where structural relaxation is required, we use fast machine-learned interatomic potentials instead of performing a new DFT relaxation for each generated structure. We further compare a coordinate-denoising control with continuous-time variance-preserving (VP) and sub-VP diffusion objectives using a randomly initialized Orb-v2 backbone. The source comparison covers seven MatBench property-prediction tasks, and an eighth task, \\texttt{mp\\_gap}, is included in the protostructure scaling study. Every synthetic-pretraining condition numerically outperforms random initialization on all seven source-comparison tasks. Scaling the protostructure corpus from 20k to 165k under a fixed 400-epoch schedule numerically matches or exceeds MP-20k pretraining on most tasks; because this protocol increases both data diversity and optimization updates, the result reflects joint data-and-compute scaling.","source_metadata":{"categories":["cs.LG","cond-mat.mtrl-sci"],"code_status":"absent"}},{"id":"preprints:2611.22098v1","kind":"preprints","source":"arXiv","title":"Every Measurement, Every Direction, All at Once: Multimodal Flow Matching for Molecules and Spectra","url":"https://arxiv.org/abs/2611.22098v1","detail_url":"/chemradar/article?u=https%3A%2F%2Farxiv.org%2Fabs%2F2611.22098v1","date":"2026-11-17T02:30:01Z","timestamp":1794882601,"categories":["Design de novo","Spectra & Analytical"],"topic_ids":["denovo","chemometrics"],"keywords":["generative models","denoising","transformer"],"matched_keywords":["generative models","denoising","transformer","Molecules","molecular"],"matched_tags":["generative_model","diffusion","transformer","generic_chem"],"doi":null,"external_id":"2611.22098v1","pdf_url":"https://arxiv.org/pdf/2611.22098v1","code_url":null,"code_host":null,"authors":["Thorben Prein","Elton Pan","Rafael Gómez-Bombarelli","Santiago Miret"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"Generative modeling in science is often data-limited, where experiments are expensive. However, each experiment may yield several coupled measurements of the same physical entity. Standard generative models often train only on a single direction, discarding supervision from the rest. We argue that each experiment contains more training signal than this asymmetry allows the model to exploit: random masking enables measurements to serve as either conditioning inputs or prediction targets, yielding multiple supervised denoising tasks from a single experiment. We investigate this premise with MOSAIC, a multimodal flow matching transformer over molecular 3D structure and UV, IR, and Raman spectra. MOSAIC is a single model that is capable of any-to-any generation: generating any subset of modalities from any disjoint subset of the others. On QM9S, MOSAIC achieves 88.8% molecular graph recovery, outperforming the strongest spectra-to-structure baseline by 23.2% in structure generation, while simultaneously outperforming a structure-to-spectra specialist model in spectral generation. Ablations show that the any-to-any objective improves structure generation over structure-only training across model sizes. Furthermore, MOSAIC supports arbitrary subsets of available measurements and enables molecular identification by matching generated spectra to experimental references.","source_metadata":{"categories":["cs.LG","physics.chem-ph"],"code_status":"absent"}},{"id":"feeds:https://thisweekincheminformatics.substack.com/p/forty-seven","kind":"feeds","source":"This Week in Cheminformatics (Substack)","title":"This Week In Cheminformatics: Issue #047","url":"https://thisweekincheminformatics.substack.com/p/forty-seven","detail_url":"/chemradar/article?u=https%3A%2F%2Fthisweekincheminformatics.substack.com%2Fp%2Fforty-seven","date":"2026-11-17T01:51:37+00:00","timestamp":1794880297,"categories":["Blog"],"topic_ids":[],"keywords":["Cheminformatics"],"matched_keywords":["Cheminformatics"],"matched_tags":["cheminformatics"],"doi":null,"external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":[],"journal":null,"publisher":null,"impact_factor":null,"abstract":"","source_metadata":{"source":"This Week in Cheminformatics (Substack)","published_utc":"2026-11-17T01:51:37+00:00","seen_at":"2026-11-17T02:20:55.012330+00:00"}},{"id":"journals:10.1039/d6dd00707d","kind":"journals","source":"Digital Discovery","title":"A single reference-energy outlier nearly eliminates Co-O chemistry from a generative model's training distribution","url":"https://doi.org/10.1039/d6dd00707d","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1039%2Fd6dd00707d","date":"2026-11-17T01:45:14Z","timestamp":1794879914,"categories":[],"topic_ids":[],"keywords":["generative model"],"matched_keywords":["generative model","chemical"],"matched_tags":["generative_model","generic_chem"],"doi":"10.1039/d6dd00707d","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Dowon Song"],"journal":"Digital Discovery","publisher":"RSC","impact_factor":5.6,"abstract":"An element-resolved audit of 44,688 oxide structures generated by MatterGen under chemical-system conditioning revealed that the aggregate composition compliance (73.27% exact match) concealed a severe chemistry-specific failure. Cobalt was omitted...","source_metadata":{"short_name":"Digital Discovery","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"journals:69f786052897a3bc500372f0791331a111638e34","kind":"journals","source":"Chemistry","title":"A Machine Learning-Based Framework for Virtual Screening of Lysine-Specific Demethylase 1 Inhibitors","url":"https://doi.org/10.3390/chemistry8110162","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.3390%2Fchemistry8110162","date":"2026-11-17T00:00:00Z","timestamp":1794873600,"categories":["Cheminformatics","Property Prediction","Docking & Screening","Free Energy & MD"],"topic_ids":["cheminformatics","property","docking","freeenergy"],"keywords":["RDKit","molecular docking","Virtual Screening","Random Forest","classification model","binding free energy","pIC50"],"matched_keywords":["RDKit","molecular docking","Virtual Screening","Random Forest","Machine Learning","classification model","binding free energy","drug","molecular","ligand","compounds","protein","pIC50"],"matched_tags":["cheminformatics_software","docking","virtual_screening","ensemble","machine_learning","ml_task","thermodynamics","generic_chem","biology","bioactivity"],"doi":"10.3390/chemistry8110162","external_id":"69f786052897a3bc500372f0791331a111638e34","pdf_url":null,"code_url":null,"code_host":null,"authors":["Lu-Yu Zhang","Yu-Ting Hao","Xiao-Fei Nan","Ling-Bo Qu","Jin-Shuai Song"],"journal":"Chemistry","publisher":null,"impact_factor":null,"abstract":"Lysine-specific demethylase 1 (LSD1) is a crucial epigenetic target closely correlated with the pathological progression of various solid tumors. In recent years, machine learning-driven virtual screening has been increasingly employed in novel drug discovery, providing promising paradigms for the rational development of LSD1 inhibitors. However, efficient screening of LSD1 inhibitors to prioritize candidates for experimental validation is still required. In this study, we propose a virtual screening framework that integrates a classification model, a drug–target affinity (DTA) prediction model (pharma-DTA), molecular docking, and binding free energy calculations. The optimal performance of the classification model is Random Forest with RDKit descriptors, which achieves a high precision of 0.930. The pharma-DTA model was established by integrating features of the ligand, pharmacophore, and the binding pocket of the protein. The model was fine-tuned on the dataset of LSD1 inhibitors, with the mean standard error of predicted pIC50 decreased to 0.5. These models were used to screen potential LSD1 inhibitors from the ZINC library, combined with molecular docking and MM-PBSA binding free energy calculations, and four top hits were identified. This virtual screening framework enables rapid identification of hit compounds for the discovery of potential LSD1 inhibitors.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:6601e237f44416c73a90e7932b11e1cbba56f986","kind":"journals","source":"Frontiers in Molecular Biosciences","title":"A microglia-derived molecular signature identifies FKBP5 as a regulator of neuroinflammatory activation in Alzheimer’s disease","url":"https://doi.org/10.3389/fmolb.2026.2003525","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.3389%2Ffmolb.2026.2003525","date":"2026-11-17T00:00:00Z","timestamp":1794873600,"categories":["Docking & Screening"],"topic_ids":["docking"],"keywords":["molecular docking","regression model"],"matched_keywords":["molecular docking","regression model","molecular","compounds"],"matched_tags":["docking","ml_task","generic_chem"],"doi":"10.3389/fmolb.2026.2003525","external_id":"6601e237f44416c73a90e7932b11e1cbba56f986","pdf_url":null,"code_url":null,"code_host":null,"authors":["Zhen-Qiang Xie","Shi-Long Han","Gui-Fang Wen"],"journal":"Frontiers in Molecular Biosciences","publisher":null,"impact_factor":null,"abstract":"Microglial dysfunction is critically involved in Alzheimer’s disease (AD) progression, yet microglia-derived molecular signatures that reflect neuropathological progression and reveal potential therapeutic targets remain incompletely understood. This study aimed to establish a microglia-based risk score model for predicting AD progression and to validate the functional role of model genes in microglial activation. Single-cell RNA sequencing datasets from AD brain tissues were analyzed to characterize disease-associated microglial subpopulations. Weighted gene co-expression network analysis (WGCNA) was performed to identify gene modules associated with Braak neuropathological stages. Microglial marker genes from advanced stage were integrated with Braak-related module genes to identify candidate genes. A least absolute shrinkage and selection operator (LASSO) regression model was constructed to generate a microglia-derived risk score. The relationship between risk score and immune microenvironment features was evaluated. FKBP5 was selected for experimental validation using an Aβ1-42-induced BV2 microglial activation model, followed by molecular docking to explore potential FKBP5-targeting compounds. Single-cell transcriptomic analysis revealed neuropathological progression in microglial states during AD progression, with enrichment of activated and immune-related microglial signatures. Integration of microglial markers and Braak-associated WGCNA modules identified 25 overlapping genes involved in immune regulation and microglial function. The LASSO-derived risk score effectively reflected AD progression and was positively associated with microglial activation and immune remodeling. Among the candidate genes, FKBP5 showed increased expression during AD progression and was closely correlated with activated microglial phenotypes. In vitro experiments demonstrated that Aβ1-42 stimulation significantly increased FKBP5 expression and promoted microglial activation via NF-κB signaling, whereas FKBP5 inhibition attenuated activation-related cytokine expression. Molecular docking analysis further suggested potential FKBP5-interacting compounds including Apatinib, Eslicarbazepine acetate, Cefadroxil, Thioisonicotinamide, and Atropine. These computationally predicted interactions warrant further experimental validation to explore their potential pharmacological relevance. We developed a microglia-derived molecular risk score that reflects AD neuropathological progression and immune microenvironment remodeling. FKBP5 was identified as a critical regulator of microglial activation and represents a potential therapeutic target for modulating neuroinflammation in AD. These findings provide new insights into microglia-mediated AD progression and establish a framework for integrating single-cell transcriptomics with experimental validation for biomarker and target discovery.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:10.1088/2632-2153/aeb9a7","kind":"journals","source":"Machine Learning: Science and Technology","title":"A provenance-aware zirconia conductivity database and Arrhenius-constrained screening of doped electrolyte candidates","url":"https://doi.org/10.1088/2632-2153/aeb9a7","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1088%2F2632-2153%2Faeb9a7","date":"2026-11-17T00:00:00+00:00","timestamp":1794873600,"categories":[],"topic_ids":[],"keywords":["density functional theory","DFT","molecular dynamics"],"matched_keywords":["neural network","density functional theory","DFT","molecular","molecular dynamics"],"matched_tags":["neural_network","dft","generic_chem","molecular_dynamics"],"doi":"10.1088/2632-2153/aeb9a7","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Qikai Zhang","Zhihao Jin","Xianfu Chen","Hao Xiong","Yan Yue","Xili Wan","Yiqun Fan","Xin Chen","Fan Zhang"],"journal":"Machine Learning: Science and Technology","publisher":"IOP","impact_factor":6.3,"abstract":"Zirconia solid electrolytes are key oxygen-ion conductors, but reported conductivity values depend strongly on dopant chemistry, processing history and measurement temperature. Here we curate a provenance-aware zirconia conductivity database of 1,351 records and use it to compare an Arrhenius-constrained physics-informed machine-learning model (PIML) with an unconstrained deep neural network (DNN). In a reserved grouped-material confirmation split, the DNN achieves the lower pooled root-mean-square error (RMSE, in log 10 σ units) for withheld composition-processing groups. The Arrhenius-constrained model remains on a comparable error scale and provides bounded effective activation-energy estimates. Its clearest benefit appears in matched-material temperature extrapolation: when trained at ≤ 800 • C and tested at ≥ 900 • C (n = 47), PIML reduces RMSE from 0.4187 to 0.3609 and raises R 2 from 0.8562 to 0.8932. Bootstrap analysis supports this record-level gain, with material-fingerprint resampling reported as a conservative uncertainty check. Genetic-algorithm (GA) screening under fixed processing and measurement conditions identifies a Sc-rich/Mg-containing zirconia region. Exploratory CHGNet molecular dynamics provides the main computational ranking check, with single-configuration oxide-reference density functional theory (DFT) retained as an auxiliary thermodynamic screen. This work provides a curated conductivity resource and a conservative screening route for doped zirconia electrolytes.","source_metadata":{"short_name":"MLST","source":"crossref","abstract_quality":"unverified"}},{"id":"journals:b33bb332362e79a695d314c7652f455f9568c2fe","kind":"journals","source":"Analytical Chemistry","title":"A Rabbit Monoclonal Antibody-Based Lateral Flow Immunoassay for Triphenyl Phosphate Enabled by Single-B-Cell-Based Two-Step Screening","url":"https://doi.org/10.1021/acs.analchem.6c05941","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1021%2Facs.analchem.6c05941","date":"2026-11-17T00:00:00Z","timestamp":1794873600,"categories":["Docking & Screening"],"topic_ids":["docking"],"keywords":["Molecular docking","shape complementarity","IC50"],"matched_keywords":["Molecular docking","shape complementarity","molecule","Molecular","IC50"],"matched_tags":["docking","shape_similarity","generic_chem","bioactivity"],"doi":"10.1021/acs.analchem.6c05941","external_id":"b33bb332362e79a695d314c7652f455f9568c2fe","pdf_url":null,"code_url":null,"code_host":null,"authors":["Tong He","Ge-Lin Liu","Peng-Hui Zhou","X. Yu","Zhan-Hui Wang","K. Wen"],"journal":"Analytical Chemistry","publisher":null,"impact_factor":null,"abstract":"Triphenyl phosphate (TPHP) is an emerging organophosphate ester contaminant associated with adverse health effects, yet immunoassay-based methods for its rapid screening remain unavailable. Rabbit monoclonal antibodies (RmAbs) are attractive recognition elements for small-molecule immunoassays, but efficient strategies for generating small-molecule-specific RmAbs are still limited. Here, we developed a single-B-cell-based two-step screening (SCTS) strategy to generate RmAbs against TPHP. The workflow combined dual fluorescence-activated cell sorting for enrichment of hapten−specific B cells with icELISA-based screening of culture supernatants to identify antibodies recognizing free TPHP. From cell sorting to functional validation, the process was completed within 7 days. Using this strategy, three RmAbs were obtained, with icELISA half-maximal inhibitory concentration (IC50) values of 0.021−0.055 ng/mL. Compared with the mouse monoclonal antibodies evaluated in this study, the RmAbs exhibited higher sensitivity and specificity, together with greater tolerance to heat, pH, ionic strength, and acetonitrile. Molecular docking suggested that the improved performance may be associated with the extended CDR-H3 of RmAb3B11, which formed a deeper and more restrictive binding cavity with higher shape complementarity and stronger steric discrimination toward TPHP. Based on RmAb3B11, a lateral flow immunoassay (LFIA) was established for rapid on-site detection of TPHP, with visual detection limits of 0.2 ng/mL in river water, 1.0 ng/g in dust, and 0.8 ng/mL in raw and packaged milk. Analysis of 40 blind samples showed good agreement between LFIA and UPLC−MS/MS results, supporting the accuracy and reliability of the method.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:43a3e6dc82c5ce7a2176640eca8e4feeca26a2bc","kind":"journals","source":"Discover Chemistry","title":"A review on the role of natural bioactive compounds in envenomation therapy","url":"https://doi.org/10.1007/s44371-026-01066-w","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1007%2Fs44371-026-01066-w","date":"2026-11-17T00:00:00Z","timestamp":1794873600,"categories":["Docking & Screening","ADMET & Safety"],"topic_ids":["docking","admet"],"keywords":["molecular docking","virtual screening","molecular dynamics"],"matched_keywords":["molecular docking","virtual screening","compounds","molecular","drug","compound","chemical","molecular dynamics","protein"],"matched_tags":["docking","virtual_screening","generic_chem","molecular_dynamics","biology"],"doi":"10.1007/s44371-026-01066-w","external_id":"43a3e6dc82c5ce7a2176640eca8e4feeca26a2bc","pdf_url":null,"code_url":null,"code_host":null,"authors":["Boma Abiye Fubara","N. M. Uzoekwe"],"journal":"Discover Chemistry","publisher":null,"impact_factor":null,"abstract":"Snakebite envenomation remains a critical neglected tropical disease, affecting approximately 5.4 million people annually and causing 81,000–138,000 deaths worldwide. Traditional antivenoms, predominantly animal-derived immunoglobulins, face significant limitations including high production costs, cold-chain dependency, hypersensitivity reactions and limited efficacy against venom-induced tissue necrosis, particularly low molecular mass toxins responsible for local tissue damage. This review synthesizes current knowledge on natural bioactive compounds—including alkaloids, flavonoids, tannins, terpenes, saponins, coumestans, sulphated polysaccharides, phenolic acids (gallic acid, ellagic acid, caffeic acid, ferulic acid, rosmarinic acid and chlorogenic acid) and endogenous human proteins—as alternative or adjunctive therapies for snakebite management. We examine their sources, mechanisms of action, structure-activity relationships and therapeutic potential against key venom toxins including phospholipase A₂ (PLA₂), snake venom metalloproteinases (SVMPs), and serine proteases. Plant-derived secondary metabolites demonstrate broad-spectrum inhibition of venom enzymes through active-site binding (flavonoids, coumestans, terpenes), metal chelation (tannins, gallic acid, quinones), protein precipitation (tannins, saponins, quinones) and antioxidant and anti-inflammatory activities. Endogenous human proteins—serum albumin, tissue inhibitors of metalloproteinases (TIMPs), C-reactive protein (CRP) and antithrombin III—offer complementary neutralisation with reduced immunogenicity, though circulating concentrations of TIMPs (50–100 ng/mL) and CRP (< 10 mg/L) are substantially below therapeutic requirements, necessitating exogenous supplementation. Notably, plant-derived compounds also inhibit catalytically inactive Lys49 PLA₂ homologues—a capability that conventional antivenoms lack—representing a significant advantage for managing venom-induced myotoxicity and edema. Synergistic combinations of endogenous proteins with phytochemicals, operating through multi-target binding, allosteric modulation, sequential neutralization, complementary pharmacokinetics and reduced immunogenicity, present a promising framework for next-generation, multi-target antivenom therapies. Recent advances in computational biology, including molecular docking, molecular dynamics simulations and AI-driven drug discovery, have accelerated the identification and optimization of plant-derived venom inhibitors as demonstrated by the identification of labdane lactone (83% inhibition at 22.7 µM) and anisic acid (100% neutralization) through virtual screening. However, despite extensive preclinical validation, no bioactive compound has yet progressed to clinical trials for snakebite envenomation—a critical translational gap that this review systematically addresses through a proposed Phase I–III clinical translation framework. This review critically evaluates the current evidence base, identifies lead candidates, compares natural products with classical inhibitors (EDTA, PMSF, benzamidine) and proposes a structured clinical translation pathway, including trial design, regulatory considerations, nanotechnology-based delivery systems, recombinant antivenoms, monoclonal antibodies and integration with WHO snakebite strategies. The review is organized into nine interconnected sections covering bioactive compound classes, sources, mechanisms, structure-activity relationships, integration with existing antivenom frameworks, ethnobotanical perspectives, challenges including chemical stability and formulation and future directions. By bridging traditional knowledge with modern pharmacology, this review provides a forward-looking framework for developing accessible, cost-effective and safer alternatives for snakebite management, contributing to Sustainable Development Goal 3 (Good Health and Well-being) and addressing a pressing global health inequity.","source_metadata":{"source":"semantic_scholar"}},{"id":"preprints:10.26434/chemrxiv.15012404/v1","kind":"preprints","source":"ChemRxiv","title":"A Revised Force Field Parameterization for Iodine-Based Halogen-Bonded Cocrystals","url":"https://doi.org/10.26434/chemrxiv.15012404/v1","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.26434%2Fchemrxiv.15012404%2Fv1","date":"2026-11-17T00:00:00Z","timestamp":1794873600,"categories":["Free Energy & MD"],"topic_ids":["freeenergy"],"keywords":["Force Field Parameterization","force field parameters","DFT","Force Field","Force fields"],"matched_keywords":["Force Field Parameterization","force field parameters","force field was demonstrated by reproducing cell parameters","DFT","Force Field","Force fields","molecular"],"matched_tags":["force_field_parameterization","dft","force_fields","generic_chem"],"doi":"10.26434/chemrxiv.15012404/v1","external_id":"10.26434/chemrxiv.15012404/v1","pdf_url":null,"code_url":null,"code_host":null,"authors":["Sibananda G. Dash","Mihails Arhangelskis"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"Halogen-bonded cocrystals have garnered attention due to their structural diversity and functional applications. Prediction of their crystal structures opens wide opportunities for computationally-driven design of functional materials based on halogen bonding. Crucial prerequisite for reliable crystal structure prediction (CSP) of halogen-bonded materials is a fast yet accurate method of relative energy ranking of the predicted structures. Force fields are preferred for CSP of molecular crystalline materials due to their low computational cost compared to periodic DFT calculations, the treatment of supramolecular interactions by heavy halogen atoms poses notable challenges. Here we report the reparameterization of a well-known exp-6 force field parameters with an emphasis on iodine containing halogen-bonded binary cocrystals. Parameterization was performed on 112 low temperature crystal structures. The multipoles derived from both the gas phase and polarizable continuum models were utilized in the calculation. The reparameterized force fields were subsequently validated against a set of 29 cocrystal structures. The accuracy of our reparameterized force field was demonstrated by reproducing cell parameters within a 3% error margin for the majority of cases. Furthermore, we achieved significant progress in the predictions of cell density and halogen bond lengths, with improvements over earlier parameterizations reaching 92%.","source_metadata":{"date_field":"posted","keywords":[]}},{"id":"preprints:10.26434/chemrxiv.15012360/v1","kind":"preprints","source":"ChemRxiv","title":"A stopping rule for completing computed phase diagrams with generative models","url":"https://doi.org/10.26434/chemrxiv.15012360/v1","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.26434%2Fchemrxiv.15012360%2Fv1","date":"2026-11-17T00:00:00Z","timestamp":1794873600,"categories":[],"topic_ids":[],"keywords":["generative models","Density functional theory"],"matched_keywords":["machine learning interatomic potential","generative models","machine learning","Density functional theory"],"matched_tags":["mlip_tools","generative_model","machine_learning","dft"],"doi":"10.26434/chemrxiv.15012360/v1","external_id":"10.26434/chemrxiv.15012360/v1","pdf_url":null,"code_url":null,"code_host":null,"authors":["Seungbae Kim","Juo Kim","Jaehwi Park","Yerim Jung","Kyoungmin Min"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"Phase diagrams constructed using known phases misjudge stability when key phases are missing. Generative models supply as many structures as needed, but how many structures to generate and which diagrams will change are unknown. We add 10,000 MatterGen structures per system, relaxed with a machine learning interatomic potential, to Materials Project (MP) data for six Li–transition-metal–O ternaries and track the hull convergence. The responses fall into two groups: Li–Ru–O, Li–Hf–O, and Li–Zr–O (70–107 MP structures) gain 4–9 stable candidates and exhibit total convex-hull shifts of 59.9–62.9 meV/atom, whereas Li–Mn–O, Li–Co–O, and Li–Ni–O (240–609) exhibit a minimal hull shift. When reduced to 70 MP structures, the latter gain a similar number of stable candidates by rediscovering removed hull phases. The hulls converge within 5,000–8,000 structures. After generating 8,000 structures, a stopping rule, tested using Li–Ir–O, decides whether more are needed. Density functional theory places seven Li–Ru–O compositions 5–83 meV/atom below the MP-only hull but rejects the potential’s false positives in well-explored systems. This gives a stopping rule and early diagnostic for incomplete phase diagrams.","source_metadata":{"date_field":"posted","keywords":[]}},{"id":"journals:10.1021/acs.jcim.6c03241","kind":"journals","source":"Journal of Chemical Information and Modeling","title":"An Integrated Coarse-Grained\nSimulation and Graph-Transformer\nFramework for the Discovery of Self-Assembling Cyclic Peptides","url":"https://doi.org/10.1021/acs.jcim.6c03241","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1021%2Facs.jcim.6c03241","date":"2026-11-17T00:00:00+00:00","timestamp":1794873600,"categories":["Cheminformatics"],"topic_ids":["cheminformatics"],"keywords":["Transformer","molecular dynamics"],"matched_keywords":["machine learning","deep learning","Transformer","molecular","molecular dynamics"],"matched_tags":["machine_learning","transformer","generic_chem","molecular_dynamics"],"doi":"10.1021/acs.jcim.6c03241","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Runqiu Ma","Shuyao Zhang","Hao Liu","Xinyu Fan","Kelin Li","Siqi Guo","Kui Zhang","Jia Jin","Fei Ye"],"journal":"Journal of Chemical Information and Modeling","publisher":"ACS","impact_factor":5.3,"abstract":"Self-assembling cyclic peptides are promising building blocks for functional biomaterials owing to their enhanced structural stability, proteolytic resistance, and diverse supramolecular architectures. However, how cyclization regulates peptide self-assembly remains poorly understood, and computational frameworks specifically developed for cyclic peptides are still lacking. Here, we demonstrate that disulfide-mediated cyclization fundamentally reshapes peptide self-assembly by significantly enhancing aggregation propensity (AP) and promoting the formation of diverse nanostructures compared with their linear counterparts, as confirmed by coarse-grained molecular dynamics (CGMD) simulations and experimental characterization. Motivated by these findings, we established an integrated computational-experimental framework for the prediction and rational design of self-assembling disulfide-cyclized peptides. A data set comprising 8,000 cyclic pentapeptides was generated using CGMD simulations, and multiple machine learning and deep learning models were systematically evaluated, among which a Graph-Transformer achieved the best predictive performance (R2 = 0.971). The optimized model was subsequently applied to high-throughput screening of 160,000 cyclic hexapeptides using a ranking strategy integrating AP and hydrophobicity, leading to the identification of multiple previously unreported self-assembling candidates. Experimental validation showed excellent agreement with model predictions. Furthermore, systematic sequence analysis revealed residue- and position-dependent assembly rules unique to disulfide-cyclized peptide systems. This work establishes a computational framework for the rapid discovery of self-assembling disulfide-cyclized peptides and provides new mechanistic insights and design principles for peptide-based functional materials.","source_metadata":{"short_name":"JCIM","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"journals:3d0e8869cd769f631f6aa262556a6fd7ef545efc","kind":"journals","source":"Nutrients","title":"Antarctic Krill Peptides Attenuate Type 2 Diabetes Mellitus-Associated Muscle Deterioration in Mice: Insights from Skeletal Muscle Metabolomics and AMPK Energy Sensing","url":"https://doi.org/10.3390/nu18223817","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.3390%2Fnu18223817","date":"2026-11-17T00:00:00Z","timestamp":1794873600,"categories":["Docking & Screening"],"topic_ids":["docking"],"keywords":["molecular docking"],"matched_keywords":["molecular docking","molecular"],"matched_tags":["docking","generic_chem"],"doi":"10.3390/nu18223817","external_id":"3d0e8869cd769f631f6aa262556a6fd7ef545efc","pdf_url":null,"code_url":null,"code_host":null,"authors":["Dong-Xue Xia","Ruo-Yao Zhang","Lei Zhai","Meng-Yao Wang","Ya-Xuan Bao","You-Shuo Wang","Hui-Lin Ma","Gui-Fang Cao","Shi-You Dai","Tian-Lin Gao"],"journal":"Nutrients","publisher":null,"impact_factor":null,"abstract":"Background: Type 2 diabetes mellitus (T2DM) induces progressive neuromuscular impairment, metabolic dysregulation, and skeletal muscle deterioration, with limited targeted nutritional countermeasures. As a promising deep-sea bioactive complex, Antarctic krill peptides (AKP) exhibit diverse metabolic-regulatory and cytoprotective properties, yet their efficacy and localized mechanisms in diabetic muscle wasting remain unestablished. Methods: Here, we investigated whether AKP attenuate T2DM-associated skeletal muscle deterioration in male C57BL/6J mice with diet- and streptozotocin-induced T2DM, using the glucokinase activator dorzagliatin as a positive control. Results: AKP was enriched in low-molecular-weight oligopeptides (<500 Da) and functional amino acids. Over the 12-week intervention, untreated diabetic mice exhibited progressive deterioration in muscular strength, fiber cross-sectional area, and tissue bioenergetics. Dietary supplementation with AKP (200 and 400 mg/kg/day) significantly alleviated body weight loss, improved glucose tolerance, reduced fasting blood glucose, and preserved forelimb grip strength. Furthermore, AKP replenished intramuscular ATP stores in soleus and tibialis anterior muscles, mitigated myofiber cross-sectional area atrophy, and normalized circulating myostatin. Untargeted metabolomics demonstrated that AKP remodeled muscular metabolic networks, enriching the tricarboxylic acid cycle, coenzyme A biosynthesis, and purine metabolism. In addition, AKP upregulated muscular p-AMPK, GLUT4, and UCP3 expressions, with molecular docking identifying 1-methylguanosine as a candidate for potential interaction with AMPK. Conclusions: Overall, AKP attenuate T2DM-associated skeletal muscle deterioration coincident with skeletal muscle bioenergetic remodeling and modulation of metabolic intermediate profiles.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:10.1038/s41467-026-79733-w","kind":"journals","source":"Nature Communications","title":"Author Correction: A machine learning model reveals expansive downregulation of ligand-receptor interactions that enhance lymphocyte infiltration in melanoma with developed resistance to immune checkpoint blockade","url":"https://doi.org/10.1038/s41467-026-79733-w","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1038%2Fs41467-026-79733-w","date":"2026-11-17T00:00:00+00:00","timestamp":1794873600,"categories":[],"topic_ids":[],"keywords":["receptor"],"matched_keywords":["machine learning","ligand","receptor"],"matched_tags":["machine_learning","generic_chem","biology"],"doi":"10.1038/s41467-026-79733-w","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Sahil Sahni","Binbin Wang","Di Wu","Saugato Rahman Dhruba","Matthew Nagy","Sushant Patkar","Ingrid Ferreira","Chi-Ping Day","Kun Wang","Eytan Ruppin"],"journal":"Nature Communications","publisher":"Nature","impact_factor":15.7,"abstract":"","source_metadata":{"short_name":"Nat. Commun.","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"journals:7c9b2a870cbd35ce5fc052a8601642c197c82edf","kind":"journals","source":"Frontiers in Behavioral Neuroscience","title":"Baihui–Zusanli electroacupuncture improves sucrose preference and lowers hippocampal S100B in mice exposed to chronic unpredictable mild stress","url":"https://doi.org/10.3389/fnbeh.2026.1973693","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.3389%2Ffnbeh.2026.1973693","date":"2026-11-17T00:00:00Z","timestamp":1794873600,"categories":["Docking & Screening"],"topic_ids":["docking"],"keywords":["molecular docking"],"matched_keywords":["molecular docking","molecular","compound","protein"],"matched_tags":["docking","generic_chem","biology"],"doi":"10.3389/fnbeh.2026.1973693","external_id":"7c9b2a870cbd35ce5fc052a8601642c197c82edf","pdf_url":null,"code_url":null,"code_host":null,"authors":["Jia-Hui Ni","Chang-Feng Cheng","Lin-Jing Wang","Chun-Ying Wang"],"journal":"Frontiers in Behavioral Neuroscience","publisher":null,"impact_factor":null,"abstract":"S100B is an astrocyte-associated protein implicated in glial stress and neuroinflammation. We examined behavioral outcomes and hippocampal S100B after electroacupuncture (EA) in mice exposed to chronic unpredictable mild stress (CUMS) and used separate computational analyses to prioritize hypotheses for investigation. Forty-eight male C57BL/6 mice were randomized to non-stressed Control, CUMS, CUMS + sham EA, and CUMS + EA groups ( n = 12 per group). After 4 weeks of CUMS, stress was discontinued during a two-week intervention. EA targeted GV20 and bilateral ST36; sham EA involved non-acupoint needling and electrical stimulation. Sucrose preference and open field tests were performed before and after intervention, followed by hippocampal S100B Western blotting ( n = 6 per group). Separate analyses combined S100B-centered reverse network pharmacology, molecular docking, and human depression blood transcriptomics. GSE52790 supported candidate discovery and nested model evaluation; a final training-selected 16-gene RF was locked for direct-transfer testing in GSE201332 and GSE98793 and for training-cohort SHAP interpretation. Paired analyses showed greater gains with EA than with CUMS in sucrose preference and open-field outcomes. Comparisons with sham EA supported additional open-field changes, whereas the sucrose-preference contrast was sensitive to the variance assumption. Hippocampal S100B/β-actin ratios were lower with EA than with CUMS or sham EA (Tukey-adjusted p < 0.0001 and p = 0.0011, respectively). Computational analyses identified 625 compound–disease overlapping targets, 1,105 differentially expressed genes, and 16 common candidate genes after four-set intersection. Mean nested outer AUC was 0.8217; external AUCs were 0.7250 in GSE201332 and 0.5499 in GSE98793, but MDD sensitivity was zero at the fixed 0.5 threshold in both external cohorts. Baihui–Zusanli EA was associated with improved sucrose preference and lower hippocampal S100B in CUMS-exposed mice. Separate transcriptomic analyses identified exploratory candidates for further functional investigation.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:10.1021/acs.jcim.6c03066","kind":"journals","source":"Journal of Chemical Information and Modeling","title":"Benchmark-Optimized Pipeline Identifies Experimentally\nValidated Antifreeze Peptides in Antarctic Krill","url":"https://doi.org/10.1021/acs.jcim.6c03066","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1021%2Facs.jcim.6c03066","date":"2026-11-17T00:00:00+00:00","timestamp":1794873600,"categories":["Cheminformatics"],"topic_ids":["cheminformatics"],"keywords":[],"matched_keywords":[],"matched_tags":[],"doi":"10.1021/acs.jcim.6c03066","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Jingxuan Qiu","Dong Zhang","Longfei Feng","Junting Xie","Xiaochuan Luo","Richa Ha","Jiahao Ma","Sihan Lu","Mingxin Song","Tianyi Qiu","Xin Wang"],"journal":"Journal of Chemical Information and Modeling","publisher":"ACS","impact_factor":5.3,"abstract":"Antifreeze proteins (AFPs) offer promising prospects as biocryoprotectants, yet their discovery remains a great challenge due to the lack of efficient large-scale screening tools. Here, we present a benchmark-optimized, multimodel computational pipeline that integrates property evaluation and structural assessment for high-throughput AFP identification. Through systematic benchmarking on independent testing data set, two optimal strategies were identified: the two-model-based Strategy 8 achieved the highest accuracy of 0.82, while the stricter four-model-based Strategy 18 attained the highest precision of 0.97, making it optimal for initial large-scale screening. Furthermore, this pipeline was applied to screen 153,359 Antarctic krill sequences and discovered 13 novel antifreeze peptides. Dynamic ice shaping assays revealed that four out of 13 peptides significantly inhibited ice crystal growth. Notably, Peptide 10 (SPHSWANTYQQCK) could inhibit ice growth and recrystallization. This study establishes a robust framework for large-scale AFP identification and provides a set of experimentally validated Antarctic krill peptides with strong potential as novel biobased cryoprotectants.","source_metadata":{"short_name":"JCIM","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"journals:574a522c4a8f33fc8dc46c1c7780a8482555f97d","kind":"journals","source":"Critical reviews in toxicology","title":"Beyond animal testing: integrated AI-PBPK-MPS framework for human-relevant toxicology","url":"https://doi.org/10.1080/10408444.2026.2725349","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1080%2F10408444.2026.2725349","date":"2026-11-17T00:00:00Z","timestamp":1794873600,"categories":["Property Prediction","ADMET & Safety"],"topic_ids":["property","admet"],"keywords":["QSAR","graph neural networks"],"matched_keywords":["QSAR","neural networks","graph neural networks","deep learning","drug"],"matched_tags":["qsar_qspr","neural_network","gnn","machine_learning","generic_chem"],"doi":"10.1080/10408444.2026.2725349","external_id":"574a522c4a8f33fc8dc46c1c7780a8482555f97d","pdf_url":null,"code_url":null,"code_host":null,"authors":["P. Wal","Jyotsana Dwivedi","Kanika Pandey","Sudarshana Borah","K. Sharma","Abid Khan","Hanish Singh Jayasingh Chellammal","Amin Gasmi"],"journal":"Critical reviews in toxicology","publisher":null,"impact_factor":null,"abstract":"NAMs are revolutionizing safety assessment and bridging the substantial disconnect between traditional animal toxicology studies and human exposure within the pharmaceutical industry. AI predictive toxicology uses graph neural networks, deep learning, explainable AI, PBPK, and IVIVE modeling for human-dose predictions, and micro-physiological and organ-on-a-chip systems that replicate specific physiological states. The development of NAMs has been applied to FDA initiatives, such as the Predictive Toxicology Roadmap, ISTAND MIDD, and CiPA, to demonstrate how integrated NAM systems optimized at early drug development stages provide valuable information on mechanistic toxicities related to humans and accelerate drug development. AI-QSAR models, PBPK simulations, and organ-on-a-chip technologies are increasingly integrated with discovery and early-stage preclinical workflows in different sectors. Standardization and reproducibility across platforms, transparency of AI models, multiomics integration, and global regulator convergence are required to improve the power of AI. This review is the first to provide a unified mechanistic synthesis of all three pillars of NAMs, AI-based predictive toxicology, PBPK-IVIVE modeling, and microphysiological systems (MPS), while earlier reviews covered them independently. We suggest a cyclical NAM ecosystem in which AI predictions seek translation to human pharmacokinetics, with iterative mechanistic validation via MPS, solving current challenges of translation from animals to humans. This review is distinct from others in that it presents NAM technologies as an integrated, human-relevant system rather than disparate technologies, and thus offers regulators a roadmap.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:1a5e4c09d39a6fe747d467df9d633b578ba32647","kind":"journals","source":"Molecules","title":"Blood–Brain Barrier Permeability Prediction: Machine Learning, Graph Neural Networks, and Protonation-State Robustness","url":"https://doi.org/10.3390/molecules31224164","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.3390%2Fmolecules31224164","date":"2026-11-17T00:00:00Z","timestamp":1794873600,"categories":["Cheminformatics"],"topic_ids":["cheminformatics"],"keywords":["Morgan fingerprint","RDKit","Graph Neural Networks","GNNs","GNN","XGBoost","Permeability Prediction","molecular feature","molecular features"],"matched_keywords":["Morgan fingerprint","RDKit","Neural Networks","Graph Neural Networks","GNNs","GNN","Machine Learning","ML","deep learning","XGBoost","Permeability Prediction","molecular feature","molecular features","drug","molecular","compounds","compound"],"matched_tags":["molecular_fingerprints","cheminformatics_software","neural_network","gnn","machine_learning","ensemble","permeability","molecular_descriptor","generic_chem"],"doi":"10.3390/molecules31224164","external_id":"1a5e4c09d39a6fe747d467df9d633b578ba32647","pdf_url":null,"code_url":null,"code_host":null,"authors":["Bo-Cheng Xie","Dong-Jing Yang","Xu-Hang Dai","Peng-Wei Xiao","Chao Yang"],"journal":"Molecules","publisher":null,"impact_factor":null,"abstract":"Blood–brain barrier permeability (BBBP) prediction is important for central nervous system (CNS) drug discovery, yet the relative performance of traditional molecular feature-based models and graph neural networks (GNNs), as well as the sensitivity of BBBP predictions to molecular protonation states, remains insufficiently characterized. Here, we curated the public OmniBBBP dataset through structural deduplication, label-conflict resolution, and removal of compounds overlapping with an external benchmark, yielding 8836 compounds for model development. Four traditional machine learning (ML) models and three GNN architectures (GCN, GAT, and PNA) were evaluated using repeated similarity-cluster cross-validation (CV). XGBoost combining 173 RDKit descriptors with a 1024-bit Morgan fingerprint achieved the strongest overall performance, with accuracies of 0.893 and 0.860 and Matthews correlation coefficients (MCCs) of 0.765 and 0.720 on the fixed test set and 200-compound external benchmark, respectively. Its benchmark performance was comparable to that reported for the deep learning model BBBP-Atlas, whereas PNA showed complementary performance among compounds with low structural similarity to the training set. To investigate protonation-state sensitivity, benchmark compounds were expanded into 962 protomeric structures, and XGBoost was retrained using a protomer-expanded training set while retaining the same molecular features and hyperparameters. Protomer-expanded training increased benchmark accuracy from 0.830 to 0.850 and MCC from 0.660 to 0.701, indicating improved robustness to alternative protonation-state representations. Overall, these results demonstrate that descriptor- and fingerprint-based XGBoost provides a competitive BBBP classification baseline under the present data and evaluation conditions and that incorporating alternative protomeric representations during training can further improve prediction robustness.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:acd250e931bc85bfc35d6daafe82040201d80199","kind":"journals","source":"The Journal of chemical physics","title":"Calculation of interfacial tensions between solid, liquid, and vapor phases in metals by molecular simulations","url":"https://doi.org/10.1063/5.0351570","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1063%2F5.0351570","date":"2026-11-17T00:00:00Z","timestamp":1794873600,"categories":["Free Energy & MD","ML Potentials"],"topic_ids":["freeenergy","potentials"],"keywords":["free energy perturbation","thermodynamic integration","free energy","molecular dynamics"],"matched_keywords":["free energy perturbation","thermodynamic integration","free energy","molecular","molecular dynamics"],"matched_tags":["free_energy_calcs","thermodynamics","generic_chem","molecular_dynamics"],"doi":"10.1063/5.0351570","external_id":"acd250e931bc85bfc35d6daafe82040201d80199","pdf_url":null,"code_url":null,"code_host":null,"authors":["M. Canducci","E. Bourasseau","J. Tranchida","P. Malfreyt"],"journal":"The Journal of chemical physics","publisher":null,"impact_factor":null,"abstract":"The interfacial free energy is a key determinant of material behavior, governing phenomena such as wetting, adhesion, nucleation, and phase coexistence. Despite its importance, experimental measurements remain challenging, particularly for liquid metals and solid-liquid interfaces under extreme conditions. Classical molecular dynamics (CMD) provides a versatile alternative, enabling atomistic evaluation of interfacial properties beyond experimental reach. In this work, we systematically compute solid-vapor, solid-liquid, and liquid-vapor interfacial free energies across multiple metals, applying and comparing established atomistic methods, including both the mechanical and thermodynamic routes, free energy perturbation, and thermodynamic integration. This unified framework allows cross-validation between methods and interfaces, revealing consistent trends and highlighting the influence of interatomic potentials on accuracy. Our results demonstrate the robustness of CMD for interfacial predictions and emphasize the importance of careful potential validation. These findings provide a coherent picture of metallic interfacial thermodynamics, with implications for mesoscale modeling, materials design, and high-temperature applications.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:10.1093/bib/bbag642","kind":"journals","source":"Briefings in Bioinformatics","title":"CausalMut: a causal deep learning framework for enzyme mutational effect prediction","url":"https://doi.org/10.1093/bib/bbag642","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1093%2Fbib%2Fbbag642","date":"2026-11-17T00:00:00+00:00","timestamp":1794873600,"categories":[],"topic_ids":[],"keywords":["enzyme"],"matched_keywords":["deep learning","enzyme"],"matched_tags":["machine_learning","biology"],"doi":"10.1093/bib/bbag642","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Wenli Gu","Fanxu Zeng","Xiaoyu You","Huilei Yu","Qingchao Jiang"],"journal":"Briefings in Bioinformatics","publisher":"Oxford","impact_factor":7.7,"abstract":"Accurate prediction of enzyme mutational effects on catalytic kinetics is fundamental to data-driven enzyme engineering, yet existing deep learning models are fundamentally constrained by biases inherent in the data generation process. Experimental datasets are conditioned on prevailing experimental choices rather than uniform mutation interventions, which introduces spurious correlations that compromise model generalization. To address this challenge, CausalMut is proposed as a causality-guided deep learning framework that disentangles genuine mutational causal effects from confounding biases. The data generation mechanism is first formalized via a structural causal graph, revealing the experimental environment as a confounder that induces distributional imbalance across mutation types. Building on this analysis, CausalMut integrates two components: a Multi-scale Interaction Enhanced Network (MIENet), which captures complex interactions among the wild-type enzyme, substrate, and mutant without 3D structural information; and a Joint Focal-IPS optimization strategy, which corrects distributional imbalances at both the amino-acid and mutant levels. Comprehensive experiments on three real-world enzyme kinetic datasets ($k_{\\text{cat}}$, $K_{m}$, and $k_{\\text{cat}}/K_{m}$) against five state-of-the-art baselines demonstrate average improvements of 16.24% in root mean square error and 22.19% in $R^{2}$. CausalMut maintains consistent superiority under three biologically motivated distribution-shift settings, and case studies further demonstrate its ability to identify high-activity variants within a fixed screening budget, validating its robustness and practical utility.","source_metadata":{"short_name":"Brief. Bioinform.","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"journals:10.1021/acs.jcim.6c03068","kind":"journals","source":"Journal of Chemical Information and Modeling","title":"Co-Folding Models\nLocalize Enzyme Active Sites but\nFail to Accurately Predict Enzyme–Substrate Complex Structures","url":"https://doi.org/10.1021/acs.jcim.6c03068","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1021%2Facs.jcim.6c03068","date":"2026-11-17T00:00:00+00:00","timestamp":1794873600,"categories":["Docking & Screening","Targets & Structures"],"topic_ids":["docking","targets"],"keywords":["AlphaFold3","AutoDock Vina","molecular dynamics","Enzyme"],"matched_keywords":["AlphaFold3","AutoDock Vina","ML","molecular","molecular dynamics","Enzyme"],"matched_tags":["ml_protein_ligand","docking_software","machine_learning","generic_chem","molecular_dynamics","biology"],"doi":"10.1021/acs.jcim.6c03068","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Daniil Lepikhov","Ariane Nunes-Alves"],"journal":"Journal of Chemical Information and Modeling","publisher":"ACS","impact_factor":5.3,"abstract":"The accurate prediction of enzyme–substrate complex structures is challenging due to the transient and dynamic nature of these interactions. We evaluated the performance of eight machine–learning (ML) docking tools and the physics–based method AutoDock Vina as a reference in the prediction of enzyme–substrate complex structures using a curated data set of 210 diverse, non–redundant enzyme–substrate complexes, equally divided into three classes: metal–free enzymes (A), and enzymes with metal ions outside (B) or inside (C) the active site. All ML–based docking tools had low geometric accuracy and physical plausibility, with fewer than 10% hit rates at the stricter 2 Å threshold for substrate root-mean-square deviation. While no ML-based tools were able to outperform Vina in modeling physically plausible complexes, co–folding tools such as AlphaFold3 and Boltz–2 excelled at active–site localization, and many algorithms performed better for class C enzymes, and worse for class B enzymes. Surprisingly, DynamicBind, which was trained on less data with fewer parameters compared to other models, matched co-folding methods in active-site elucidation. Our results show that enzyme–substrate docking can be affected by the presence of a metal ion, and suggest that better performances for complex structure prediction may be achieved with data-centric strategies including molecular dynamics simulations rather than parameter-heavy ML architectures.","source_metadata":{"short_name":"JCIM","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"journals:5edcc9433445c204e4d63bf71a303a4a2664bd70","kind":"journals","source":"Journal of Chemical Information\nand Modeling","title":"Co-Folding Models\nLocalize Enzyme Active Sites but\nFail to Accurately Predict Enzyme–Substrate Complex Structures","url":"https://doi.org/10.1021/acs.jcim.6c03068","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1021%2Facs.jcim.6c03068","date":"2026-11-17T00:00:00Z","timestamp":1794873600,"categories":[],"topic_ids":[],"keywords":["AlphaFold3","AutoDock Vina","molecular dynamics","Enzyme"],"matched_keywords":["AlphaFold3","AutoDock Vina","ML","molecular","molecular dynamics","Enzyme"],"matched_tags":["ml_protein_ligand","docking_software","machine_learning","generic_chem","molecular_dynamics","biology"],"doi":"10.1021/acs.jcim.6c03068","external_id":"5edcc9433445c204e4d63bf71a303a4a2664bd70","pdf_url":null,"code_url":null,"code_host":null,"authors":["Daniil Lepikhov","Ariane Nunes-Alves"],"journal":"Journal of Chemical Information\nand Modeling","publisher":null,"impact_factor":null,"abstract":"The accurate prediction of enzyme–substrate complex structures is challenging due to the transient and dynamic nature of these interactions. We evaluated the performance of eight machine–learning (ML) docking tools and the physics–based method AutoDock Vina as a reference in the prediction of enzyme–substrate complex structures using a curated data set of 210 diverse, non–redundant enzyme–substrate complexes, equally divided into three classes: metal–free enzymes (A), and enzymes with metal ions outside (B) or inside (C) the active site. All ML–based docking tools had low geometric accuracy and physical plausibility, with fewer than 10% hit rates at the stricter 2 Å threshold for substrate root-mean-square deviation. While no ML-based tools were able to outperform Vina in modeling physically plausible complexes, co–folding tools such as AlphaFold3 and Boltz–2 excelled at active–site localization, and many algorithms performed better for class C enzymes, and worse for class B enzymes. Surprisingly, DynamicBind, which was trained on less data with fewer parameters compared to other models, matched co-folding methods in active-site elucidation. Our results show that enzyme–substrate docking can be affected by the presence of a metal ion, and suggest that better performances for complex structure prediction may be achieved with data-centric strategies including molecular dynamics simulations rather than parameter-heavy ML architectures.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:db68dd5485e56c8aad83021dd734ac2ed49bce03","kind":"journals","source":"Discover Food","title":"Comparative phytochemical profiling and in silico evaluation of α-amylase inhibitory potential of functional biscuits fortified with Ficus exasperata leaves and Sphenostylis stenocarpa seeds","url":"https://doi.org/10.1007/s44187-026-01347-4","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1007%2Fs44187-026-01347-4","date":"2026-11-17T00:00:00Z","timestamp":1794873600,"categories":["Docking & Screening","ADMET & Safety","Free Energy & MD"],"topic_ids":["docking","admet","freeenergy"],"keywords":["molecular docking","binding free energy","ADMET","pharmacokinetic","pIC50","receptor"],"matched_keywords":["molecular docking","binding free energy","ADMET","pharmacokinetic","pIC50","compounds","molecular","receptor"],"matched_tags":["docking","thermodynamics","admet","bioactivity","generic_chem","biology"],"doi":"10.1007/s44187-026-01347-4","external_id":"db68dd5485e56c8aad83021dd734ac2ed49bce03","pdf_url":null,"code_url":null,"code_host":null,"authors":["S. Shodehinde","Lateef Bello","O. V. Awelewa","D. O. Nwankwo","Samuel A. Oginni","O. V. Awojulu","S. Yusuf","S. O. Olubode","Tobiloba S. Omoboyowa","Precious J. Akinnadeju","Oyindamola F. Olususi","H. Sanusi","Ayomide S. Adetula","O. Balogun","P. Ajayi","Islamiat K. Lanloye","Testimony O. Kazeem"],"journal":"Discover Food","publisher":null,"impact_factor":null,"abstract":"Functional foods enriched with plant-derived bioactives have gained considerable attention as dietary sources of health-promoting phytochemicals. This study aimed to comparatively characterize the bioactives retained in biscuits fortified with fermented versus unfermented Ficus exasperata leaves and Sphenostylis stenocarpa seeds, and to evaluate the α-amylase inhibitory potential of the HPLC-identified compounds using computational approaches. Two biscuit formulations (B-SS20-uFE40 and B-SS20-fFE40) containing 20 g Sphenostylis stenocarpa seed flour and 40 g of either unfermented or fermented Ficus exasperata leaf powder were prepared. Biscuit extracts were evaluated for in vitro α-amylase inhibitory activity, while secondary metabolites were profiled by HPLC. The identified compounds were subjected to molecular docking against human pancreatic α-amylase (PDB ID: 4GQR), followed by MMGBSA calculations, induced-fit docking (IFD), AutoQSAR prediction, and ADMET evaluation. B-SS20-uFE40 formulation showed 47.22% α-amylase inhibition while B-SS20-fFE40 showed 42.16% inhibition. HPLC detected fourteen bioactive compounds. Most bioactives predominated in the B-SS20-uFE40 formulation, whereas caffeic acid, myricetin and p-coumaric acid were enriched in the B-SS20-fFE40 formulation. Molecular docking identified baicalin, quercetin, kaempferol, and caffeic acid (− 8.803 to − 6.642 kcal/mol) as promising α-amylase-binding compounds. MMGBSA and AutoQSAR prediction showed baicalin possessed favourable binding free energy (-53.08 kcal/mol) and predicted pIC50 = 5.356. IFD suggested possible receptor-flexible binding poses, ADMET analysis predicted pharmacokinetic and toxicity-related properties that require experimental confirmation. The fortified biscuits retained detectable phenolic and flavonoid constituents after baking. Fermentation modulated metabolite composition. The computational findings provide a basis for further experimental validation of these biscuits as functional food products.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:b1966838c3c78a4d0cf2f21092ddcfa9e4dbf041","kind":"journals","source":"Zeitschrift für Naturforschung C","title":"Coriaceolide A, a new calanolide-type coumarin isolated from the leaves of Calophyllum coriaceum M.R. Hend. & Wyatt-Sm. with antimicrobial, antibiofilm, and molecular docking studies","url":"https://doi.org/10.1515/znc-2026-0207","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1515%2Fznc-2026-0207","date":"2026-11-17T00:00:00Z","timestamp":1794873600,"categories":["Docking & Screening"],"topic_ids":["docking"],"keywords":["molecular docking"],"matched_keywords":["molecular docking","molecular","compounds","Compound","drug"],"matched_tags":["docking","generic_chem"],"doi":"10.1515/znc-2026-0207","external_id":"b1966838c3c78a4d0cf2f21092ddcfa9e4dbf041","pdf_url":null,"code_url":null,"code_host":null,"authors":["N. M. Mohd Zaini","Wan Salleh","N. Ab Ghani","Nurulfazlina Edayah Rasol","A. Salihu","M. Arzmi","Sherali Kuziev","F. Khasanov","F. Eshboev","J. Nyemb","L. Nahar","S. D. Sarker"],"journal":"Zeitschrift für Naturforschung C","publisher":null,"impact_factor":null,"abstract":"Phytochemical investigation of the leaves of Calophyllum coriaceum afforded a new calanolide-type coumarin, coriaceolide A (1), along with seven known compounds, including β-amyrin, calanolide A, calanolide B, calanolide E1, friedelin, lupeol, and β-sitosterol. The structure of coriaceolide A (1) was elucidated using comprehensive spectroscopic techniques (NMR, IR, UV, and MS). Compound 1 exhibited antimicrobial activity against Candida albicans (MIC 125 μg/mL) and Streptococcus mutans (MIC 62.5 μg/mL), with both fungicidal and bactericidal effects at these concentrations. In antibiofilm assays, compound 1 demonstrated strong inhibition (83.45 ± 0.18 %), particularly against S. mutans. Molecular docking against lanosterol 14α-demethylase (ERG11) and glucosyltransferase B (GtfB) showed favorable binding affinities (−7.1 and −8.3 kcal/mol, respectively), supporting the biological results. These findings highlight calanolide-type coumarins as promising scaffolds for antimicrobial and antibiofilm drug development.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:9ff8291d9b6635bcdff812a383a8120ed0bb77a7","kind":"journals","source":"Current Issues in Molecular Biology","title":"Cross-Disease Integration Identifies Inflammation-Related Candidate Biomarkers for Atherosclerosis and Candidate Therapeutic Compounds Using Bioinformatics, Machine Learning, and Molecular Dynamics Simulations","url":"https://doi.org/10.3390/cimb48111242","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.3390%2Fcimb48111242","date":"2026-11-17T00:00:00Z","timestamp":1794873600,"categories":["Docking & Screening"],"topic_ids":["docking"],"keywords":["molecular docking","computational screening","Molecular Dynamics","binding affinity"],"matched_keywords":["molecular docking","computational screening","Machine Learning","protein","Compounds","Molecular","molecule","compound","Molecular Dynamics","binding affinity"],"matched_tags":["docking","virtual_screening","machine_learning","biology","generic_chem","molecular_dynamics","bioactivity"],"doi":"10.3390/cimb48111242","external_id":"9ff8291d9b6635bcdff812a383a8120ed0bb77a7","pdf_url":null,"code_url":null,"code_host":null,"authors":["Mei-Dong Si","Xue-Mei Ma","Ya-Ni Yang","X. Tian","Yao-Dong Yang","Ru Zhou"],"journal":"Current Issues in Molecular Biology","publisher":null,"impact_factor":null,"abstract":"Objective: This study aimed to establish a translational research framework integrating cross-disease computational screening with experimental validation. Using Helicobacter pylori (H. pylori) infection as a cross-disease inflammatory comparator, we systematically identified atherosclerosis-associated hub genes through integrated bioinformatics and machine-learning analyses, evaluated their potential diagnostic value in independent datasets, and validated their protein expression in a mouse model of atherosclerosis (AS). Methods: GEO gene-expression datasets were analyzed using differential expression analysis, WGCNA, functional enrichment, and machine-learning algorithms. CIBERSORT was used to assess immune-cell infiltration and gene–immune associations in AS and H. pylori infection. miRNA–mRNA–TF and TF–hub gene networks were constructed. Candidate compounds were predicted using DGIdb, followed by molecular docking, molecular dynamics simulations, and in vivo validation. Results: Integration of differential expression analysis and WGCNA identified 656 candidate genes associated with AS and H. pylori infection. Eleven key genes were subsequently identified using MCODE and cytoHubba. Machine-learning analysis further identified CX3CR1 and TNF as potential diagnostic biomarker candidates. ROC curve analysis demonstrated that CX3CR1 exhibited more consistent discriminatory performance across datasets, whereas TNF showed considerable variability in diagnostic performance, particularly in the independent validation cohorts. These findings should be interpreted cautiously given the limited sample sizes of some cohorts. Immune infiltration analysis revealed associations between CX3CR1, TNF, and immune dysregulation related to H. pylori infection and AS. Additionally, miRNA–mRNA–TF regulatory network analysis suggested that these two genes may be involved in inflammatory signaling pathways associated with disease progression. Molecular docking and molecular dynamics simulations indicated that the selected small-molecule compound exhibited favorable binding affinity for the target protein and formed a structurally stable complex. In vivo experiments demonstrated that CX3CR1 and TNF protein levels were significantly increased in atherosclerotic lesions in an AS mouse model. Conclusions: Using a two-stage strategy combining cross-disease computational screening with experimental validation, we identified CX3CR1 and TNF as candidate inflammation-related biomarkers associated with AS and H. pylori infection. CX3CR1 demonstrated more consistent diagnostic potential than TNF across the analyzed datasets. In vivo validation revealed significant upregulation of both proteins in atherosclerotic lesions, supporting their potential involvement in AS pathogenesis. Collectively, these findings suggest that CX3CR1 may represent a promising candidate diagnostic biomarker for atherosclerosis, whereas the diagnostic performance and pathological significance of TNF require further investigation.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:10.1038/s41598-026-78838-y","kind":"journals","source":"Scientific Reports","title":"Design and evaluation of indole-based hydrazones targeting heparanase with cytotoxic activity in human breast cancer cells","url":"https://doi.org/10.1038/s41598-026-78838-y","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1038%2Fs41598-026-78838-y","date":"2026-11-17T00:00:00+00:00","timestamp":1794873600,"categories":["Docking & Screening","ADMET & Safety"],"topic_ids":["docking","admet"],"keywords":["molecular docking","DFT","ADME","molecular dynamics"],"matched_keywords":["molecular docking","DFT","ADME","drug","molecular","compounds","compound","molecular dynamics"],"matched_tags":["docking","dft","admet","generic_chem","molecular_dynamics"],"doi":"10.1038/s41598-026-78838-y","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["M. Padmapriya","R. Manjunath","Erik W. Thompson","Alan K. L. Toh","Narasimha M. Beeraka","Mamatha Sindhuvalli Kempasiddegowda","Peter E. Lobie","Vladimir N. Nikolenko","Basappa Basappa","Rajeev K. Sinha","Neha S. Gandhi","Santosh L. Gaonkar"],"journal":"Scientific Reports","publisher":"Nature","impact_factor":null,"abstract":"Heparanase, an endoglycosidase implicated in tumour growth, invasion, and metastasis,​ remains an attractive target for anticancer drug discovery. In this study, a series of indole-based hydrazones was designed to target heparanase with potential anticancer inhibition in breast cancer cells. Structure-based analyses, including molecular docking and molecular dynamics simulations, suggested that several derivatives can interact with residues in the heparanase binding region, including recurrent contacts involving GLU 225, GLN 270, HIS 296, TYR 298, and TYR 348. In silico ADME profiling indicated generally acceptable drug-like properties for the designed series. The designed compounds were synthesized and characterized by FTIR, 1 H NMR, 13 C NMR, and HRMS, and conformational features were further examined by DFT studies. Among the synthesized compounds evaluated for heparanase inhibition using the HTRF assay, derivative E showed the highest apparent inhibitory response; however, its activity remained substantially weaker than that of the reference inhibitor PG545 and lacked a clear dose-dependent trend, indicating only modest inhibitory potential. Furthermore, cytotoxicity testing in MCF-7 breast cancer cells identified compound A as the most active compound, with an IC 50 value of 2.078 µM. Overall, these results identify indole-based hydrazones as preliminary scaffolds for further optimization toward heparanase-related anticancer applications.","source_metadata":{"short_name":"Sci. Rep.","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"journals:10.1021/acs.jafc.6c12335","kind":"journals","source":"Journal of Agricultural and Food Chemistry","title":"Design, Synthesis, and Antifungal Evaluation of Disulfide-Containing\nStrobilurin Derivatives as Potential Fungicides","url":"https://doi.org/10.1021/acs.jafc.6c12335","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1021%2Facs.jafc.6c12335","date":"2026-11-17T00:00:00+00:00","timestamp":1794873600,"categories":["Property Prediction"],"topic_ids":["property"],"keywords":["QSAR","EC50"],"matched_keywords":["QSAR","compound","EC50"],"matched_tags":["qsar_qspr","generic_chem","bioactivity"],"doi":"10.1021/acs.jafc.6c12335","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Xiaofeng Liu","Shuqing Ma","Qingqiang Tian","Shaochen Li","Yahui Li"],"journal":"Journal of Agricultural and Food Chemistry","publisher":"ACS","impact_factor":null,"abstract":"To address the persistent challenge of fungicide resistance in crop protection, a series of disulfide-containing strobilurin derivatives were designed, synthesized, and evaluated for their antifungal activity. Guided by three-dimensional quantitative structure–activity relationship (3D-QSAR) analysis, compound 5ag exhibited potent activity against Curvularia lunata (EC50 = 4.32 mg/L), comparable to that of the commercial fungicide kresoxim-methyl (KM, EC50 = 9.60 mg/L). In vivo, 5ag at 200 mg/L showed curative and protective efficacies of 76.6% and 81.0% against rice leaf spot disease, respectively, compared with 59.5% and 68.2% for KM, while it showed no obvious phytotoxicity to rice under the tested conditions. Scanning electron microscopy (SEM) revealed pronounced hyphal deformation, including curvature and abnormal swelling, following treatment with 5ag. These findings demonstrate that disulfide incorporation can expand the structural diversity of strobilurin derivatives and identify 5ag as a promising lead for further antifungal optimization.","source_metadata":{"short_name":"J. Agric. Food Chem.","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"journals:4c9c7c6fc18c8d10d625b836018e1f2e56c9f15f","kind":"journals","source":"RSC advances","title":"Development of novel benzo[d][1,2,3]triazole-acetohydrazone derivatives as potential therapeutic agents: synthesis, SAR study, inhibition of carbohydrate metabolism enzymes, and aldose reductase supported by molecular modelling simulation","url":"https://doi.org/10.1039/d6ra06631c","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1039%2Fd6ra06631c","date":"2026-11-17T00:00:00Z","timestamp":1794873600,"categories":["Docking & Screening","ADMET & Safety"],"topic_ids":["docking","admet"],"keywords":["molecular docking","ADME","molecular dynamics","IC50"],"matched_keywords":["molecular docking","ADME","molecular","compounds","Compound","molecule","drug","molecular dynamics","IC50"],"matched_tags":["docking","admet","generic_chem","molecular_dynamics","bioactivity"],"doi":"10.1039/d6ra06631c","external_id":"4c9c7c6fc18c8d10d625b836018e1f2e56c9f15f","pdf_url":null,"code_url":null,"code_host":null,"authors":["Mohamed El-Naggar","Mohamed E. Belasy","Mohga Basseem","W. Aboulthana","Ahmed T. Elhendawy","A. Ragab"],"journal":"RSC advances","publisher":null,"impact_factor":null,"abstract":"Diabetes prevalence is on the rise worldwide, which makes it more urgent to develop safe, effective, and multi-targeted inhibitors. Herein, we describe the synthesis of novel bioactive benzo[d][1,2,3]triazole-acetohydrazone derivatives 4-12. These compounds were prepared via condensation reactions between benzotriazole-carbohydrazide and various formyl benzylidene derivatives, 4-formyl-1,3-diarylpyrazole, and cyclic ketones (ninhydrin and 2,3-dioxoindole), all of which are linked through an acetohydrazone moiety. The structures of derivatives 4-12 were confirmed using FT-IR, NMR (1H/13C{1H}) spectroscopy, elemental analysis, and HPLC. The synthesized analogues were assessed for their ability to mitigate hyperglycemia by inhibiting carbohydrate-metabolizing enzymes (α-amylase and α-glucosidase) and suppressing aldose reductase in the polyol pathway. Compound 11 exhibited particularly notable efficacy, demonstrating IC50 values of 6.08 ± 0.06 µg mL-1 for α-amylase and 5.01 ± 0.09 µg mL-1 for α-glucosidase. These values are comparable to those of acarbose (IC50 = 5.54 ± 0.04 µg mL-1 for α-amylase and IC50 = 4.53 ± 0.06 µg mL-1 for α-glucosidase). SAR analysis indicated that the presence of two carbonyl groups and a hydrophobic aromatic ring is crucial for synergistic activity in compound 11. Furthermore, compounds 6-12 displayed significant aldose reductase inhibitory activity, with IC50 values below 10 µg mL-1, and compound 11 demonstrated the most potent activity, with an IC50 value of 4.07 ± 0.10 µg mL-1, which closely approached that of quercetin (IC50 = 3.68 ± 0.11 µg mL-1). Consequently, compound 11 holds promise as a lead molecule for developing novel therapeutic agents targeting α-amylase, α-glucosidase, and aldose reductase for diabetes management. Furthermore, molecular docking and molecular dynamics simulations (100 ns) elucidated stable interactions between compound 11 and the catalytic pockets of both α-amylase and α-glucosidase, primarily mediated by hydrogen bonding and hydrophobic interactions. The ADME analysis revealed an acceptable physicochemical profile, characterized by good aqueous solubility and a minimal predicted risk of CYP-mediated drug-drug interaction. Finally, these findings indicate that benzotriazole derivative 11 may serve as a promising scaffold for further optimization toward a multi-target anti-diabetic drug development agent.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:10.1021/acsomega.6c09237","kind":"journals","source":"ACS Omega","title":"DFT Study of NH3 Adsorption and Decomposition\non Zr(0001)","url":"https://doi.org/10.1021/acsomega.6c09237","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1021%2Facsomega.6c09237","date":"2026-11-17T00:00:00+00:00","timestamp":1794873600,"categories":[],"topic_ids":[],"keywords":["DFT","density functional theory"],"matched_keywords":["ML","DFT","density functional theory"],"matched_tags":["machine_learning","dft"],"doi":"10.1021/acsomega.6c09237","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Weronika Walkosz","Joseph Maret","Sofia Strupovets","Juan C. Garcia","Hakim Iddir"],"journal":"ACS Omega","publisher":"ACS","impact_factor":null,"abstract":"Ammonia adsorption and decomposition to hydrogen and nitrogen is studied on the Zr(0001) surface using density functional theory within the generalized gradient approximation and a supercell approach at a coverage of 0.25 ML. Different adsorption sites are considered for NH3 and its dehydrogenated species. We find that ammonia adsorbs readily and decomposes to atomic nitrogen and hydrogen through an exothermic pathway with relatively small kinetic barriers. The NH decomposition step is identified as rate-limiting. Nitrogen strongly favors subsurface incorporation over surface adsorption, whereas hydrogen remains on the surface. The reaction energetics and kinetic barriers obtained in this work are compared with previous studies of NH3 adsorption and decomposition via sequential hydrogen abstraction on ZrB2(0001).","source_metadata":{"short_name":"ACS Omega","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"journals:5c2fcfe0c7bbb66ae7e2ec1c9d5de31485309ba8","kind":"journals","source":"Antioxidants","title":"Effects of Amentoflavone-Standardized Juniperus oxycedrus Fruit Extract on Systemic Redox and Hemodynamic Responses in Rat Endotoxemia","url":"https://doi.org/10.3390/antiox15111494","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.3390%2Fantiox15111494","date":"2026-11-17T00:00:00Z","timestamp":1794873600,"categories":["Docking & Screening"],"topic_ids":["docking"],"keywords":["molecular docking"],"matched_keywords":["molecular docking","molecular"],"matched_tags":["docking","generic_chem"],"doi":"10.3390/antiox15111494","external_id":"5c2fcfe0c7bbb66ae7e2ec1c9d5de31485309ba8","pdf_url":null,"code_url":null,"code_host":null,"authors":["Ü. Bozok","B. Ergene","G. Böyük Özcan","Fatma Uysal Cinar","D. Özen","B. Bayraktar"],"journal":"Antioxidants","publisher":null,"impact_factor":null,"abstract":"Endotoxemia-associated coagulopathy integrates inflammation, oxidative stress, circulatory dysfunction, fibrin turnover, and organ injury. This study evaluated an amentoflavone-standardized methanolic Juniperus oxycedrus fruit extract in lipopolysaccharide (LPS)-induced endotoxemia in rats and explored candidate targets using network pharmacology and molecular docking. Male Wistar rats were assigned to Control, LPS, LPS + aspirin (200 mg/kg), LPS + J. oxycedrus 100 mg/kg, or LPS + J. oxycedrus 200 mg/kg groups (n = 10/group). HPLC-DAD, thrombo-inflammatory biomarkers, systemic redox indices, antioxidant defenses, hemodynamic and physiological variables, hepatic and renal injury markers, apelin, lung wet/dry ratio, and descriptive histopathology were integrated. The extract contained approximately 0.41% amentoflavone. Network analysis identified 139 core targets, and docking predicted favorable binding poses for amentoflavone–MMP-9 (−9.5 kcal/mol) and luteolin–COX-2 (−9.4 kcal/mol). LPS increased D-dimer, oxidative injury, IL-6, lactate, ALT, AST, and creatinine and reduced MAP, core temperature, apelin, and antioxidant defenses. Both extract doses reduced D-dimer; 200 mg/kg produced the broadest numerical systemic response, lowering MDA by 51.6%, TOS by 50.4%, IL-6 by 65.1%, lactate by 59.2%, ALT by 54.3%, AST by 52.4%, and creatinine by 50.8% versus LPS, while increasing MAP by 46.3%, SOD by 87.4%, CAT by 86.8%, GPx-1 by 102.4%, TAS by 80.4%, and apelin by 132.8%. TxA2, tissue factor, BAL cytokines, and lung wet/dry ratio did not show a consistent extract-related improvement. The findings support a multi-domain association between J. oxycedrus treatment and improved systemic redox, perfusion, inflammatory, and organ-injury profiles, while the docking results remain hypothesis-generating and do not establish target engagement. Responses were endpoint-specific; uniform dose dependence, statistically confirmed between-dose superiority, comprehensive anticoagulation, and definitive pulmonary protection were not established.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:10.1021/acscatal.6c06716","kind":"journals","source":"ACS Catalysis","title":"Evolution-Guided Engineering of a Process-Compatible Thymidine Phosphorylase for the Scalable Cascade Biosynthesis of 2′/3′-Functionalized Purine Nucleoside Analogs","url":"https://doi.org/10.1021/acscatal.6c06716","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1021%2Facscatal.6c06716","date":"2026-11-17T00:00:00+00:00","timestamp":1794873600,"categories":[],"topic_ids":[],"keywords":["virtual screening","Molecular dynamics","enzyme"],"matched_keywords":["virtual screening","Molecular","Molecular dynamics","enzyme"],"matched_tags":["virtual_screening","generic_chem","molecular_dynamics","biology"],"doi":"10.1021/acscatal.6c06716","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Gaofei Liu","Jialing Wang","Fan Yang","Yuhan Sun","Fuxue Chen","Wang Xi","Xingcheng Shao","Yanbin Lu","Bin Wu"],"journal":"ACS Catalysis","publisher":"ACS","impact_factor":13.1,"abstract":"2′/3′-Functionalized nucleoside analogs are valuable therapeutic agents and essential building blocks for nucleic acid-based medicines; however, their scalable biocatalytic synthesis remains challenging due to insufficient enzyme robustness under process-relevant conditions. Here, we developed an ancestral sequence reconstruction (ASR)-guided engineering strategy integrated with kcat/Km-based virtual screening to generate a process-compatible thymidine phosphorylase (TP) variant for scalable nucleoside biosynthesis. Among the reconstructed candidates, ASR114 exhibited an optimal balance between soluble expression, catalytic efficiency, and thermostability, displaying an apparent half-life of > 386 h at 50 °C compared with 7.7 h for wild-type EcTP. Biochemical characterization demonstrated that ASR114 enhanced catalytic efficiency toward multiple 2′/3′-functionalized pyrimidine nucleoside donors, with up to a 5.4-fold improvement in kcat/Km values. Molecular dynamics simulations suggested that the enhanced catalytic performance of ASR114 was associated with reduced structural flexibility, optimized substrate interactions, and alleviated conformational strain during substrate recognition. By integrating ASR114, EcPNP, and EcADA into a multienzyme cascade system, we established a crude enzyme-based platform for the synthesis of diverse 2′/3′-functionalized purine nucleoside analogs. The optimized process was successfully scaled to 120 L, enabling kilogram-scale synthesis of eight nucleoside analogs with isolated yields of up to 80.3%. This study establishes an evolution-guided framework for engineering process-compatible biocatalysts by integrating ancestral enzyme engineering, computational prioritization, and process-oriented biocatalysis for the scalable synthesis of complex nucleoside analogs.","source_metadata":{"short_name":"ACS Catal.","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"journals:68d4a1319f017e02dff0d6695626c594b42e73fc","kind":"journals","source":"International Journal of Molecular Sciences","title":"Exploring Potential Links Among Gut Microbiota, Butyrate, and JUN in Lung Adenocarcinoma: A Bioinformatics-Based Integrative Multi-Omics Analysis","url":"https://doi.org/10.3390/ijms272210312","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.3390%2Fijms272210312","date":"2026-11-17T00:00:00Z","timestamp":1794873600,"categories":["Docking & Screening"],"topic_ids":["docking"],"keywords":["Molecular docking","molecular dynamics"],"matched_keywords":["Molecular docking","Molecular","molecular dynamics","protein"],"matched_tags":["docking","generic_chem","molecular_dynamics","biology"],"doi":"10.3390/ijms272210312","external_id":"68d4a1319f017e02dff0d6695626c594b42e73fc","pdf_url":null,"code_url":null,"code_host":null,"authors":["Bo-Xuan Qiu","Yu-Peng Cai","Yi-Kuan Cai","Hua-Jie Li","Jing-Jing Zhang","Xiao-He Huang","Zhi-Zhi Wang","Hao-Fei Wang","Zhi-Ming Chen"],"journal":"International Journal of Molecular Sciences","publisher":null,"impact_factor":null,"abstract":"Lung adenocarcinoma (LUAD) is highly heterogeneous, and the mechanisms linking gut microbiota-derived metabolites to tumor-associated transcriptional regulation remain unclear. We integrated two paired LUAD transcriptomic datasets, GSE40419 and GSE32863, to perform differential expression, functional enrichment, protein–protein interaction, and transcription factor activity analyses. Candidate microbial metabolites and targets were obtained from gutMGene, Similarity Ensemble Approach, and SwissTargetPrediction, while Mendelian randomization (MR) was used to evaluate associations between genetically predicted gut microbial abundance and LUAD risk. Molecular docking and a 100 ns molecular dynamics simulation were conducted to assess interactions between candidate metabolites and Jun Proto-Oncogene (JUN)-containing DNA-bound complexes. We identified 1271 and 2519 differentially expressed genes in GSE40419 and GSE32863, respectively, including 912 overlapping genes. These genes were enriched mainly in cell cycle, immune and inflammatory, angiogenic, and tumor microenvironment-related pathways. Fifteen robust transcription factors were identified, with consistent suppression of the JUN-regulated transcriptional program. MR analysis revealed 10 microbial taxa associated with LUAD after false discovery rate correction. A genetically predicted higher abundance of Roseburia inulinivorans was associated with a lower LUAD risk (OR = 0.843, 95% CI: 0.756–0.940). Database integration linked R. inulinivorans to butyrate and JUN. Butyrate showed the most favorable docking score with the c-Jun–DNA complex (−4.5 kcal/mol), and molecular dynamics simulation indicated a relatively stable interaction. Collectively, these findings support a putative R. inulinivorans–butyrate–JUN regulatory axis that links gut microbial metabolism to transcriptional dysregulation in LUAD.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:10.1021/acscentsci.6c01736","kind":"journals","source":"ACS Central Science","title":"Firefly-Geni:\nAutomated Iteration-Guided Generative\nDiscovery of Thermally Activated Delayed Fluorescence Molecules","url":"https://doi.org/10.1021/acscentsci.6c01736","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1021%2Facscentsci.6c01736","date":"2026-11-17T00:00:00+00:00","timestamp":1794873600,"categories":["Design de novo","LLMs & Agents"],"topic_ids":["denovo","llm"],"keywords":["molecular generation","LLM","property prediction"],"matched_keywords":["molecular generation","LLM","property prediction","Molecules","molecular","chemical"],"matched_tags":["de_novo_design","LLMs","ml_task","generic_chem"],"doi":"10.1021/acscentsci.6c01736","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Boyuan Zhang","Xuerui Fang","Di Yang","Yuanyue Wang","Yunfei Liu","Yumin Xia","Qiang Zhu","Haibo Yu","Meifang Zhu","Jin Wen"],"journal":"ACS Central Science","publisher":"ACS","impact_factor":10.4,"abstract":"The efficient discovery of thermally activated delayed fluorescence (TADF) molecules remains heavily dependent on expert-driven design rules, while the vast structure–property knowledge accumulated in the literature stays largely fragmented and underutilized. Here, we present Firefly-Geni, an automated and interpretable active-evolution framework that transforms dispersed literature knowledge into structurally novel TADF candidates through a closed-loop workflow. Firefly-Geni integrates large-language-model (LLM)-assisted data extraction, multitask property prediction, conditional molecular generation, and theoretical evaluation. Its data-mining module converts unstructured literature into a curated database of experimental property entries. A dual-molecular-graph multitask predictor and a conditional generator are integrated with quantum-chemical feedback, enabling iterative refinement toward structural novelty. From 100,000 molecules generated under target conditions of a singlet–triplet energy gap ΔEST = 0.05 eV and a synthetic accessibility (SA) score of 2.5, Firefly-Geni identifies 100 structurally novel candidates computationally predicted to support reverse intersystem crossing and to be synthetically tractable, spanning diverse emission wavelengths. One synthesized candidate exhibited an experimental ΔEST of 0.019 eV and a film photoluminescence quantum yield (PLQY) of 84% under nitrogen, broadly consistent with the predicted small gap and high PLQY. Interpretability analyses further reveal chemically meaningful structure–property relationships and generation preferences, confirming that the candidates encode physically relevant low-ΔEST design motifs. This work offers a closed-loop, explainable strategy for actively evolving TADF candidates from literature knowledge, providing a generalizable route for data-driven molecular discovery.","source_metadata":{"short_name":"ACS Cent. Sci.","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"journals:d6f02ddb9aa73a6cf03281dc585078138bc06ff7","kind":"journals","source":"Natural product research","title":"Fisetin attenuates TBHP-induced senescence-associated phenotypes in human gingival fibroblasts and is associated with reduced p53/p21 expression","url":"https://doi.org/10.1080/14786419.2026.2764767","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1080%2F14786419.2026.2764767","date":"2026-11-17T00:00:00Z","timestamp":1794873600,"categories":["Docking & Screening"],"topic_ids":["docking"],"keywords":["molecular docking"],"matched_keywords":["molecular docking","molecular","protein"],"matched_tags":["docking","generic_chem","biology"],"doi":"10.1080/14786419.2026.2764767","external_id":"d6f02ddb9aa73a6cf03281dc585078138bc06ff7","pdf_url":null,"code_url":null,"code_host":null,"authors":["Min-Shan Chen","Jian-Bo Sun","Kai-Yin Chen"],"journal":"Natural product research","publisher":null,"impact_factor":null,"abstract":"Fisetin has potential anti-ageing activity, but its effects on human gingival fibroblast (HGF) senescence remain unclear. This study investigated fisetin using network pharmacology, molecular docking, and in vitro assays. Network analysis predicted fisetin-related targets enriched in cellular senescence, PI3K-Akt signalling, extracellular matrix remodelling, and inflammation pathways, with TP53, AKT1, and matrix metalloproteinases as potential key nodes. Docking suggested favourable binding of fisetin to core targets, particularly TP53. In tert-butyl hydroperoxide-induced HGFs, fisetin improved migration, reduced senescence-associated beta-galactosidase positivity, and decreased p53 and p21 protein expression. Fisetin also reduced MMP1 and MMP9 mRNA levels while restoring COL1A1 and COL3A1 mRNA expression. In a separate lipopolysaccharide-stimulated inflammation model, fisetin reduced interleukin-6 secretion. These findings suggest that fisetin attenuates selected senescence-associated phenotypes in HGFs, accompanied by reduced p53/p21 expression and improved extracellular matrix-related marker profiles, warranting further validation in tissue-level, animal, and clinical studies in future work, although causal mediation was not established.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:084cca50b90705131e90ca2846adce6e29bf28d4","kind":"journals","source":"Journal of the science of food and agriculture","title":"Flavor improvement of soy sauce driven by process optimization: analysis of key aroma compounds and molecular docking study","url":"https://doi.org/10.1002/jsfa.71225","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1002%2Fjsfa.71225","date":"2026-11-17T00:00:00Z","timestamp":1794873600,"categories":["Docking & Screening"],"topic_ids":["docking"],"keywords":["molecular docking"],"matched_keywords":["molecular docking","compounds","molecular"],"matched_tags":["docking","generic_chem"],"doi":"10.1002/jsfa.71225","external_id":"084cca50b90705131e90ca2846adce6e29bf28d4","pdf_url":null,"code_url":null,"code_host":null,"authors":["Yi-Xue Xiao","Shi-Hua Qi","Hao Jiang","Li-Hua Hou","Shou-Peng Wan"],"journal":"Journal of the science of food and agriculture","publisher":null,"impact_factor":null,"abstract":"","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:10.1021/acs.jcim.6c02114","kind":"journals","source":"Journal of Chemical Information and Modeling","title":"FlowBack-Adjoint\n                    : Physics-Aware and Energy-Guided\nConditional Flow-Matching for Protein Backmapping","url":"https://doi.org/10.1021/acs.jcim.6c02114","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1021%2Facs.jcim.6c02114","date":"2026-11-17T00:00:00+00:00","timestamp":1794873600,"categories":["Property Prediction","Design de novo"],"topic_ids":["property","denovo"],"keywords":["generative model","force field","molecular dynamics"],"matched_keywords":["generative model","force field","molecular","molecular dynamics","Protein"],"matched_tags":["generative_model","force_fields","generic_chem","molecular_dynamics","biology"],"doi":"10.1021/acs.jcim.6c02114","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Alex Berlaga","Michael S. Jones","Curt Waltmann","Gregory A. Voth","Andrew L. Ferguson"],"journal":"Journal of Chemical Information and Modeling","publisher":"ACS","impact_factor":5.3,"abstract":"Coarse-grained (CG) molecular models of proteins can substantially increase the time and length scales accessible to molecular dynamics simulations of proteins, but recovery of accurate all-atom (AA) ensembles from CG simulation trajectories can be essential for exposing molecular mechanisms and for the calculation of physical properties requiring atomistic detail. The recently reported deep generative model FlowBack restores AA detail to protein Cα traces using a flow-matching architecture and demonstrates state-of-the-art performance in backmapping of AA structural ensembles (J. Chem. Inf. Model. 65, 2, 672–692, 2025). Training, however, is performed exclusively on structural data, and in the absence of any awareness of interatomic energies or forces, small structural imperfections can produce high-energy backmapped structures. In this work, we introduce FlowBack-Adjoint as a lightweight enhancement that upgrades the pretrained FlowBack model through a one-time, physics-aware post-training pass. Auxiliary contributions to the flow introduce physical awareness of bond lengths, bond angles, and Lennard-Jones interactions, and gradients of a molecular mechanics force field energy are incorporated via adjoint matching to steer FlowBack-Adjoint toward low-energy configurations without compromising structural fidelity or diversity. In benchmark tests against FlowBack, FlowBack-Adjoint lowers single-point energies by a median of ∼30 kcal/mol·residue, reduces errors in bond lengths by >95%, eliminates >97% of molecular clashes, maintains excellent diversity of the AA configurational ensemble, and produces configurations capable of initializing stable all-atom molecular dynamics simulations. In applications to large protein–protein interfaces, including peptide–major histocompatibility complexes (peptide–MHCs) and HIV capsid assemblies, the model improves agreement with molecular dynamics-derived contact frequencies and increases the recall of experimentally observed interactions. We propose FlowBack-Adjoint as an accurate and efficient physics-aware deep generative model for AA backmapping from Cα traces generated by structure prediction tools, coarse-grained simulations, or low-resolution experimental data.","source_metadata":{"short_name":"JCIM","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"journals:619dcf2970ed3d08dc64c56dc974b823e245b596","kind":"journals","source":"Current opinion in structural biology","title":"From static structures to conformational ensembles: Artificial intelligence for protein sampling and dynamics","url":"https://doi.org/10.1016/j.sbi.2026.103438","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1016%2Fj.sbi.2026.103438","date":"2026-11-17T00:00:00Z","timestamp":1794873600,"categories":[],"topic_ids":[],"keywords":["molecular dynamics"],"matched_keywords":["machine learning","ML","drug","molecular","molecular dynamics","protein"],"matched_tags":["machine_learning","generic_chem","molecular_dynamics","biology"],"doi":"10.1016/j.sbi.2026.103438","external_id":"619dcf2970ed3d08dc64c56dc974b823e245b596","pdf_url":null,"code_url":null,"code_host":null,"authors":["Aldo S. Pasos-Trejo","Klara Bonneau","Andrea Guljas","Ana Molina-Taborda","Cecilia Clementi"],"journal":"Current opinion in structural biology","publisher":null,"impact_factor":null,"abstract":"After sparking a revolution in protein structure prediction, machine learning (ML), and artificial intelligence (AI) are now reshaping the modeling of protein conformational ensembles and dynamics. A fast, scalable, and accurate representation of protein conformational ensembles could enable genome-scale studies of protein function, with broad applications in drug discovery and protein engineering. The challenges associated with this goal, however, are far greater than those involved in single structure prediction because of the broad timescales spanned by protein motion, the complex relationship between sequence and dynamics, and the scarcity of datasets with experimental measurements. Despite these difficulties, many promising solutions have emerged in recent years, ranging from accelerated molecular dynamics to direct sampling of the conformational distribution, as well as some hybrid methods. Here, we review recent AI-powered approaches to tackle protein dynamics and conformational sampling, and outline their current limitations.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:a83d7162362432cd1b1ce4f5881e48765166ff03","kind":"journals","source":"Journal of Chemical Education","title":"Generative-AI-Assisted\nDevelopment of Browser-Based\nDocking Notebooks: A ChatGPT “Vibe Coding” Methodology\nfor Chemistry and Pharmacy Education","url":"https://doi.org/10.1021/acs.jchemed.6c00083","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1021%2Facs.jchemed.6c00083","date":"2026-11-17T00:00:00Z","timestamp":1794873600,"categories":["Cheminformatics","Docking & Screening","LLMs & Agents"],"topic_ids":["cheminformatics","docking","llm"],"keywords":["SMILES","Molecular docking","ChatGPT","receptor"],"matched_keywords":["SMILES","Molecular docking","ChatGPT","Molecular","ligand","ligands","receptor"],"matched_tags":["molecular_notation","docking","LLMs","generic_chem","biology"],"doi":"10.1021/acs.jchemed.6c00083","external_id":"a83d7162362432cd1b1ce4f5881e48765166ff03","pdf_url":null,"code_url":null,"code_host":null,"authors":["Ahmed M. Sayed"],"journal":"Journal of Chemical Education","publisher":null,"impact_factor":null,"abstract":"Molecular docking is widely taught in medicinal chemistry and pharmacy programs, yet student participation is frequently limited by software installation barriers, platform incompatibilities, and uneven access to computational resources. This report describes a generative-AI-assisted (“vibe coding”) methodology in which ChatGPT is used to iteratively draft, refine, and debug Google Colab notebooks that deliver an install-free docking experience through a web browser. Because Colab executes Python in the browser and requires no local setup, the workflow can be run on laptops or mobile devices, supporting equitable access for students. Two notebooks are reported: (i) single-ligand docking to a single receptor with SMILES input, pH-aware structure preparation, autodetection of a docking box from a co-crystallized ligand when present, and student-editable sampling parameters (grid, exhaustiveness, poses, verbosity, seed) and (ii) a library-screening notebook that accepts a SMILES list, prepares ligands in parallel with pH control, executes batch docking, and exports ranked results for analysis. The central educational contribution is not only the notebook itself but a reproducible prompt → code → test loop that enables instructors to construct browser-deployable computational notebooks, which students with minimal coding experience can then run and modify at the parameter level while practicing verification, scientific reasoning, and responsible AI disclosure.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:10.1021/acs.jcim.6c03060","kind":"journals","source":"Journal of Chemical Information and Modeling","title":"HiFiPL-Bench: A\nFidelity-First and Bias-Aware Benchmark\nfor Diagnosing Protein–Ligand Scoring Functions","url":"https://doi.org/10.1021/acs.jcim.6c03060","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1021%2Facs.jcim.6c03060","date":"2026-11-17T00:00:00+00:00","timestamp":1794873600,"categories":["Docking & Screening"],"topic_ids":["docking"],"keywords":[],"matched_keywords":[],"matched_tags":["biology","generic_chem"],"doi":"10.1021/acs.jcim.6c03060","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Mengguang Liu","Xiaodong Ren","Pengyao Jin","Likang Zhai","Rugui Xiao","Yali Zhang"],"journal":"Journal of Chemical Information and Modeling","publisher":"ACS","impact_factor":5.3,"abstract":"Accurate prediction of protein–ligand binding affinities is fundamental to structure-based drug design, yet existing benchmarks are often limited by outdated structural coverage, uncertain local structural fidelity, and physicochemical or protocol-dependent bias. Here, we present HiFiPL-Bench, a fidelity-first and bias-aware benchmark curated from RCSB PDB and BindingDB through hierarchical filtering, manual electron-density validation, affinity verification, and property-matched decoy generation. Across nine representative scoring functions, HiFiPL-Bench revealed reduced performance in scoring, ranking, and most docking evaluations relative to CASF-2016, whereas a controlled screening-like task using computationally generated property-matched decoys produced model-specific gains and losses that altered relative rankings rather than uniformly lowering performance. Model-specific overlap and deoverlap analyses further showed that documented development-data exposure differed substantially among methods and that deoverlapped sensitivity analyses affected models and evaluation tasks nonuniformly. Together, these findings indicate that performance estimates remain sensitive to starting-structure curation, task-specific preparation, data set composition, decoy design, benchmark protocol, and recoverable development-data exposure. HiFiPL-Bench therefore provides a reusable diagnostic framework for evaluating scoring-function performance in a structurally curated and bias-aware benchmark setting.","source_metadata":{"short_name":"JCIM","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"journals:10.1021/acsomega.6c06715","kind":"journals","source":"ACS Omega","title":"In Silico Discovery\nof Novel Topoisomerase IIα\nInhibitors with Enhanced Antiproliferative Activity in Breast Cancer","url":"https://doi.org/10.1021/acsomega.6c06715","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1021%2Facsomega.6c06715","date":"2026-11-17T00:00:00+00:00","timestamp":1794873600,"categories":["ADMET & Safety"],"topic_ids":["admet"],"keywords":["PubChem","ADMET","Glide SP and XP docking","Molecular dynamics"],"matched_keywords":["PubChem","ADMET","Glide SP and XP docking","drug","molecules","compounds","ligands","Molecular","Molecular dynamics"],"matched_tags":["chem_benchmarks","admet","docking_software_ambigous","generic_chem","molecular_dynamics"],"doi":"10.1021/acsomega.6c06715","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Anju Sebastian","Satabdi Debroy","Anwesha Saikia","Basavarajeshwari Police Patil","Krishnaprasad Baby","Venkata Narayanan Naranammalpuram Sundaram","Deepak Samuel","Yogendra Nayak","Duraippandi Palanimuthu"],"journal":"ACS Omega","publisher":"ACS","impact_factor":null,"abstract":"Human topoisomerase IIα (TOP2A) is frequently overexpressed in aggressive cancers, making it an attractive therapeutic target. Herein, we employed an integrated in silico and experimental approach to identify novel TOP2A inhibitors. A library of 669,069 filtered, drug-like bioactive small molecules from PubChem was screened against human TOP2A (PDB ID: 5GWK) using Schrödinger Maestro with Glide SP and XP docking, yielding 202 compounds with docking scores superior to etoposide. Subsequent filtering for PAINS, ADMET, and selectivity assessment against other human type I and type II topoisomerases yielded 21 ligands. Docking against multidomain TOP2A (PDB ID: 6ZY8) revealed preferential binding at the DNA-gate over the ATP-binding N-gate region, with L8 (pitavastatin) achieving the highest docking score. Based on their docking score, selectivity, and commercial availability, five ligands were selected for experimental evaluation. A cell viability assay showed that pitavastatin and L21 (PubChem CID: 7369491) exhibited potent cytotoxicity against MDA-MB-231 and MCF-7 breast cancer cells, respectively, with reduced toxicity against normal HEK293 cells. Both ligands induced S-phase arrest and late apoptosis in MDA-MB-231 cells, with pitavastatin outperforming etoposide. The in vitro plasmid relaxation assay confirmed that all five ligands inhibited TOP2A activity consistent with the poison mechanism, while pitavastatin exhibited a nonintercalative mode of binding to DNA. Molecular dynamics simulations for pitavastatin supported stable binding within the TOP2A active site. These findings identify pitavastatin as a promising repurposed candidate for TOP2A-targeted therapy, particularly against the triple-negative breast cancer cell line MDA-MB-231.","source_metadata":{"short_name":"ACS Omega","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"journals:32228dad1f1c3571564719787d1aa84408a52213","kind":"journals","source":"Drug development and industrial pharmacy","title":"Integrating AI-Guided Design with Experimental Validation for Niosome-Based Drug Delivery Systems","url":"https://doi.org/10.1080/03639045.2026.2761534","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1080%2F03639045.2026.2761534","date":"2026-11-17T00:00:00Z","timestamp":1794873600,"categories":["LLMs & Agents"],"topic_ids":["llm"],"keywords":["XGBoost"],"matched_keywords":["XGBoost","machine learning","Drug","drugs"],"matched_tags":["ensemble","machine_learning","generic_chem"],"doi":"10.1080/03639045.2026.2761534","external_id":"32228dad1f1c3571564719787d1aa84408a52213","pdf_url":null,"code_url":null,"code_host":null,"authors":["Y. Y. Salem","Joud Hourani","Ranim Al Rfai","S. Abd El-Gawwad","Hamzah Dabool","Tasneem Zarzour","Rana M. F. Sammour"],"journal":"Drug development and industrial pharmacy","publisher":null,"impact_factor":null,"abstract":"OBJECTIVE To develop and experimentally validate an explainable artificial intelligence (AI) framework that extracts formulation rules for niosomal drug delivery systems from published data and delivers them through a decision-support platform. SIGNIFICANCE Niosome performance depends on many interacting formulation and process variables, and optimization still relies largely on trial-and-error experimentation. Converting existing formulation knowledge into interpretable, experimentally verified rules could make early-stage development faster and less resource-intensive. METHODS A database of 33 optimized niosomal formulations (approximately 30 drugs, BCS classes I-V) was compiled from peer-reviewed studies. Correlation analysis, XGBoost feature importance, and SHAP were used to relate formulation variables to particle size, entrapment efficiency (EE%), drug release, and release time. The extracted trends were tested experimentally using 17 diclofenac sodium-loaded niosomal formulations prepared by thin-film hydration. The dataset was expanded to 1000 samples using SMOTE to train XGBoost predictive models, which were evaluated on the original formulations and integrated with a Qwen-based large language model. RESULTS Phosphate-buffered saline as the hydration medium, Span 60, and chloroform volume were the most influential variables. Laboratory results agreed with the AI-predicted direction of effect for all tested variables: cholesterol, Span 60, chloroform volume, Tween 80, and hydration medium. The predictive models achieved R2 values of 0.97-0.99 on the original formulations, and the resulting Niosome AI Expert platform showed general agreement between predicted and reported outcomes. CONCLUSIONS Combining machine learning, explainable AI, and experimental verification can convert existing formulation knowledge into a practical decision-support tool that reduces reliance on trial-and-error niosome development.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:3b2c8df20756c5172e2f8e799210988045d6918e","kind":"journals","source":"Naunyn-Schmiedeberg's archives of pharmacology","title":"Integration of machine learning and transcriptomics analysis identifies candidate molecular associations between parthenolide and sepsis-associated acute lung injury","url":"https://doi.org/10.1007/s00210-026-06063-5","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1007%2Fs00210-026-06063-5","date":"2026-11-17T00:00:00Z","timestamp":1794873600,"categories":[],"topic_ids":[],"keywords":[],"matched_keywords":["machine learning","molecular"],"matched_tags":["machine_learning","generic_chem"],"doi":"10.1007/s00210-026-06063-5","external_id":"3b2c8df20756c5172e2f8e799210988045d6918e","pdf_url":null,"code_url":null,"code_host":null,"authors":["Tayyaba S. Ali","F. Sharopov","Adam Junka","Yan-Fang Sun"],"journal":"Naunyn-Schmiedeberg's archives of pharmacology","publisher":null,"impact_factor":null,"abstract":"","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:48972b0ee702c5d3b686ad22058cc8ee23512d5c","kind":"journals","source":"Frontiers in Immunology","title":"Integrative transcriptomic, genetic, single-cell, and experimental analyses prioritize HILPDA in preeclampsia","url":"https://doi.org/10.3389/fimmu.2026.1973156","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.3389%2Ffimmu.2026.1973156","date":"2026-11-17T00:00:00Z","timestamp":1794873600,"categories":["Docking & Screening"],"topic_ids":["docking"],"keywords":["molecular docking","XGBoost","molecular dynamics"],"matched_keywords":["molecular docking","XGBoost","molecular","compounds","molecular dynamics"],"matched_tags":["docking","ensemble","generic_chem","molecular_dynamics"],"doi":"10.3389/fimmu.2026.1973156","external_id":"48972b0ee702c5d3b686ad22058cc8ee23512d5c","pdf_url":null,"code_url":null,"code_host":null,"authors":["Zong-Yu Guan","Bo-Yu Zhang","Lin Xu","Jia Ge","Yun-Zhao Xu","Meng-Meng Sang"],"journal":"Frontiers in Immunology","publisher":null,"impact_factor":null,"abstract":"Preeclampsia (PE) is a major cause of maternal and perinatal morbidity and mortality. Its molecular drivers remain incompletely understood, which limits the development of targeted therapies. This study aimed to prioritize PE-associated candidate genes and characterize their cell type-specific relevance in the placenta. Eight placental transcriptomic datasets were integrated for differential expression analysis and weighted gene co-expression network analysis. Three gradient-boosting algorithms were used to construct PE prediction models. Five additional machine-learning methods were applied for biomarker selection. Mendelian randomization was performed using two independent PE genome-wide association studies. Single-cell RNA sequencing was used to examine cell type-specific expression, transcription factor activity, intercellular communication, trophoblast co-expression networks, virtual gene perturbation, and trophoblast subtypes. The function of the prioritized gene was evaluated in lipopolysaccharide-treated HTR-8/SVneo cells. Connectivity Map screening, molecular docking, and molecular dynamics simulations were used to identify candidate compounds. Twenty-two differentially expressed and disease-associated genes were identified. The prediction models showed good performance in the mixed validation dataset, with area under the curve values of 0.841, 0.789, and 0.825 for XGBoost, LightGBM, and GBM, respectively. HILPDA was selected by all five feature-selection methods. Mendelian randomization provided genetic evidence supporting an association between HILPDA expression and PE. Single-cell analysis showed that HILPDA was predominantly expressed in trophoblasts. HILPDA expression was associated with altered trophoblast communication, inflammatory signaling, mitochondrial and translational programs, cell adhesion, and subtype-specific regulatory networks. In LPS-stimulated HTR-8/SVneo cells, HILPDA knockdown partially restored cell migration and VEGF and PlGF expression and reduced inflammatory cytokine expression, NF-κB activation, and apoptosis-related changes. Four compounds, NVP-BEZ235, vanoxerine, NVP-TAE226, and cycloserine, showed favorable predicted interactions with HILPDA. HILPDA was identified as a PE-associated candidate by transcriptomic, machine-learning, and genetic analyses and was mainly expressed in trophoblasts at the single-cell level. In LPS-stimulated HTR-8/SVneo cells, HILPDA knockdown attenuated inflammatory and pro-apoptotic responses and partially restored cell migration and angiogenic-factor expression. These findings support further investigation of HILPDA in PE-related trophoblast dysfunction, although its underlying mechanisms and therapeutic relevance require additional validation.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:08ff9b511c46f7bc53b582c1bb495ab92995d867","kind":"journals","source":"European journal of pharmacology","title":"Macelignan attenuates LPS-induced hippocampal memory impairment by suppressing HMGB1/TLR4/NLRP3-mediated microglial pyroptosis","url":"https://doi.org/10.1016/j.ejphar.2026.179584","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1016%2Fj.ejphar.2026.179584","date":"2026-11-17T00:00:00Z","timestamp":1794873600,"categories":["Docking & Screening"],"topic_ids":["docking"],"keywords":["Molecular docking"],"matched_keywords":["Molecular docking","Molecular"],"matched_tags":["docking","generic_chem"],"doi":"10.1016/j.ejphar.2026.179584","external_id":"08ff9b511c46f7bc53b582c1bb495ab92995d867","pdf_url":null,"code_url":null,"code_host":null,"authors":["Sha Li","Xin-Ru Qiu","Xin-Ge Chu","Zheng-Yu Qi","Bin Xiao","Guang-Yu Jin","Qing-Hua Jin","Chun-Ai Cui"],"journal":"European journal of pharmacology","publisher":null,"impact_factor":null,"abstract":"Neuroinflammation is a pivotal pathological process in neurological disorders. Microglial pyroptosis, a lytic and inflammatory form of programmed cell death, critically exacerbates neuroinflammation and causes neuronal damage. Macelignan has neuroprotective properties, but its mechanism in regulating microglial pyroptosis remains unclear. This study investigated whether macelignan protects against neuroinflammation by inhibiting microglial pyroptosis via high mobility group box 1 (HMGB1). In vitro, lipopolysaccharide (LPS)-stimulated BV2 microglial cells were treated with macelignan alone or combined with glycyrrhizin (an HMGB1 inhibitor) or disulfiram (a pyroptosis inhibitor). Macelignan suppressed M1 polarization and inflammatory cytokine release, as assessed by CCK-8, ELISA, and morphological analysis. Molecular docking predicted a potential interaction between macelignan and HMGB1. Western blot and immunofluorescence showed that macelignan downregulated HMGB1/TLR4/NLRP3 signaling, inhibited NLRP3 inflammasome activation, caspase-1 cleavage, and GSDMD-mediated pyroptosis, phenocopying glycyrrhizin. In a microglia-neuron co-culture model, conditioned medium from macelignan-treated BV2 cells reduced apoptosis in hippocampal HT22 neurons. In vivo, LPS-injected C57BL/6 mice were treated with low- or high-dose macelignan or glycyrrhizin (positive control). Macelignan attenuated hippocampal pyroptosis and neuronal damage, and behavioral tests revealed improved hippocampal-dependent memory. Collectively, these findings suggest that macelignan alleviates neuroinflammation and cognitive impairment likely by suppressing the HMGB1/NLRP3 axis to inhibit microglial pyroptosis, offering a potential mechanism for its neuroprotective action.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:32ea396c16db8ac7f57b74025390cc1e99999db3","kind":"journals","source":"Pharmaceuticals","title":"Machine Learning and Structure-Based Virtual Screening for Reliability-Aware G Protein-Coupled Receptor Ligand Prioritization","url":"https://doi.org/10.3390/ph19111829","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.3390%2Fph19111829","date":"2026-11-17T00:00:00Z","timestamp":1794873600,"categories":["Property Prediction"],"topic_ids":["property"],"keywords":["ChEMBL","Virtual Screening","Receptor","GPCR","bioactivity","IC50","EC50"],"matched_keywords":["ChEMBL","Virtual Screening","Machine Learning","Protein","Receptor","GPCR","Ligand","compounds","bioactivity","IC50","EC50"],"matched_tags":["chem_benchmarks","virtual_screening","machine_learning","biology","generic_chem","bioactivity"],"doi":"10.3390/ph19111829","external_id":"32ea396c16db8ac7f57b74025390cc1e99999db3","pdf_url":null,"code_url":null,"code_host":null,"authors":["S. Mtambo","S. Makhathini","H. Kumalo","C. Ssemakalu"],"journal":"Pharmaceuticals","publisher":null,"impact_factor":null,"abstract":"Background: Predictive performance is necessary but insufficient for G protein-coupled receptor (GPCR) ligand prioritization, where heterogeneous bioactivity data, analog-series overlap, receptor-state dependence, and polypharmacology can weaken high-scoring predictions. Methods: We developed an auditable ligand- and structure-based workflow across five human GPCRs: DRD2, CB2, ADORA2A, OPRM1, and CCR5. ChEMBL Ki, IC50, and EC50 data were evaluated using scaffold-disjoint validation, external validation, Venn–Abers calibration, conformal prediction, applicability-domain analysis, explainability, and receptor-state-aware docking. Results: Deployed classifiers achieved held-out ROC-AUC values of 0.876–0.929. Under stricter temporal validation, models developed exclusively on pre-cutoff data achieved ROC-AUC values of 0.617–0.894 on the same post-cutoff scaffold-held-out compounds, indicating optimism in the original temporal stress test arising from post-cutoff analog information. De-overlapped BindingDB validation was informative for ADORA2A (n = 134) and OPRM1 (n = 49), whereas smaller DRD2, CB2, and CCR5 strata were descriptive only. Venn–Abers provided validity-aware probability intervals, although calibration metrics did not improve uniformly. Classification error increased outside the applicability domain for all targets, while regression conformal coverage decreased. Morgan fingerprints dominated predictive discrimination, whereas physicochemical descriptors recurred more frequently among top SHAP features across targets. DrugBank screening yielded 92 high-confidence, dataset-novel predictions, with structure-based filtering reducing prioritization to three evidence-classified compounds across three targets. Conclusions: Overall, the study presents a reliability-aware framework for evidence-qualified GPCR ligand prioritization rather than an experimentally confirmed candidate-discovery campaign. Prospective biochemical and functional validation, including potency and selectivity assessment where appropriate, is required before candidate-level conclusions can be established.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:09cbac98400ef4248869ed677ffd8041a8878944","kind":"journals","source":"Nanoscale","title":"Machine learning-guided discovery of TMX4@G electrocatalysts with LLM-assisted workflow automation","url":"https://doi.org/10.1039/d6nr02937j","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1039%2Fd6nr02937j","date":"2026-11-17T00:00:00Z","timestamp":1794873600,"categories":["LLMs & Agents"],"topic_ids":["llm"],"keywords":["LLM","regression model","DFT","chemical diversity"],"matched_keywords":["LLM","Machine learning","regression model","DFT","chemical diversity","ligand","ligands","chemical"],"matched_tags":["LLMs","machine_learning","ml_task","dft","molecular_diversity","generic_chem"],"doi":"10.1039/d6nr02937j","external_id":"09cbac98400ef4248869ed677ffd8041a8878944","pdf_url":null,"code_url":null,"code_host":null,"authors":["Zheng-Yan Jiang","Dong-Yu Bai","Jun-Xian Liu","Liang-Zhi Kou"],"journal":"Nanoscale","publisher":null,"impact_factor":null,"abstract":"Coordination-engineered transition-metal single-atom catalysts offer a modular platform for the hydrogen evolution reaction (HER), but their discovery is hindered by the large configurational space arising from coupled metal-ligand variations. Here we develop a symmetry-aware and uncertainty-guided workflow to accelerate the screening of TMX4@G catalysts (X = C, N, O, P or S), where a transition-metal atom is coordinated by four ligands embedded in graphene. Canonicalization of rotationally equivalent motifs reduces 17 500 nominal structures to 4900 symmetry-inequivalent candidates while preserving chemical diversity. A cascaded Gaussian process regression model encodes a physical hierarchy from atomic descriptors to electronic structure, metal-hydrogen interaction and the HER activity. Posterior uncertainty drives active learning by balancing near-thermoneutral hydrogen binding with exploration of underrepresented coordination motifs. A large language model (LLM)-assisted automation framework further integrates structure generation, VASP input preparation, job submission, output parsing and dataset updating. DFT validation shows strong agreement with model predictions (R2 = 0.898 for 24 candidates) and identifies six thermodynamically stable catalysts with ηHER < 0.15 V.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:ca2f60cf7ef141b1ecc4567df1ad72becad71640","kind":"journals","source":"Physical chemistry chemical physics","title":"Natural language-guided generation of molecular structures for single-molecule electronics","url":"https://doi.org/10.1039/d6cp03430f","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1039%2Fd6cp03430f","date":"2026-11-17T00:00:00Z","timestamp":1794873600,"categories":["Design de novo","LLMs & Agents"],"topic_ids":["denovo","llm"],"keywords":["molecular generation","LLMs","LLM"],"matched_keywords":["molecular generation","LLMs","LLM","molecular","molecule","chemical","molecules"],"matched_tags":["de_novo_design","LLMs","generic_chem"],"doi":"10.1039/d6cp03430f","external_id":"ca2f60cf7ef141b1ecc4567df1ad72becad71640","pdf_url":null,"code_url":null,"code_host":null,"authors":["Yi-Heng Zhao","Bing-Qian Cheng","Bai-Lin Gao","Zi-Yang Wang","Sai-Sai Yuan","Zhi-Chao Pan"],"journal":"Physical chemistry chemical physics : PCCP","publisher":null,"impact_factor":null,"abstract":"As molecular science advances toward autonomous scientific discovery, enabling AI models to directly generate valid molecular structures from natural language descriptions has emerged as a critical requirement. While large language models (LLMs) have demonstrated the ability to map discrete chemical space into continuous semantic representation for general molecular generation, their application to single-molecule electronics (SME) remains largely unexplored. This is primarily due to the specific topological requirements of the field, such as the necessity of an anchor-backbone-anchor motif for electrode connection. In this work, we propose SingMolT5, a fine-tuned molecular LLM for natural language-guided generation of single molecules. We systematically evaluated the structural generation performance using both in-context learning (ICL) and fine-tuning strategies based on a carefully curated domain-specific instruction-tuning dataset. Results show that while ICL can capture basic molecular syntax, only the fine-tuning model can reliably enforce the anchor-backbone-anchor motif required for single-molecule measurements. More importantly, SingMolT5 achieved high structural prediction accuracy, along with strong generalization to unseen structural variations not included in the training dataset. This work provides a proof-of-concept demonstration for a text-to-structure workflow in SME, serving as a foundational step toward integrating natural language models into future autonomous workflows for single-molecule research.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:a4b3c76197dd2ab4b62f9be0269201a9656b2478","kind":"journals","source":"Frontiers in Public Health","title":"Nonylphenol and octylphenol in prostate cancer: bioinformatics identification of ABCC4/CDK4 and functional verification for nonylphenol","url":"https://doi.org/10.3389/fpubh.2026.2007850","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.3389%2Ffpubh.2026.2007850","date":"2026-11-17T00:00:00Z","timestamp":1794873600,"categories":["Docking & Screening"],"topic_ids":["docking"],"keywords":["molecular docking","chemicals"],"matched_keywords":["molecular docking","machine learning","chemicals","molecular"],"matched_tags":["docking","machine_learning","generic_chem"],"doi":"10.3389/fpubh.2026.2007850","external_id":"a4b3c76197dd2ab4b62f9be0269201a9656b2478","pdf_url":null,"code_url":null,"code_host":null,"authors":["Yong Zhu","Zi-Jian Tan","Zi-Yong Wang","Chong-Jian Zhang","Ying Bi","Yu Bai","Hong Yang","Hai-Yang Jiang"],"journal":"Frontiers in Public Health","publisher":null,"impact_factor":null,"abstract":"Endocrine-disrupting chemicals (EDCs) like nonylphenol (NP) and octylphenol (OP) are linked to prostate cancer (PCa) risk, but molecular mechanisms remain unclear. This work aimed to dissect how NP and OP drive PCa pathogenesis. Bulk and single-cell transcriptomics were combined with network toxicology, Mendelian randomization (MR), machine learning and molecular docking. Candidate genes were screened by intersection analysis, and causal relationships were evaluated via MR. Multiple machine learning algorithms identified robust biomarkers, which were further validated in external cohorts. Single-cell transcriptome profiling, functional enrichment and molecular docking were performed, followed by in vitro cellular assays. We identified 29 overlapping PCa-related targets of NP and OP, among which CDK4, AR, CDKN2A and ABCC4 were prioritized. MR suggested potential causal associations for ABCC4 (OR=1.02, p=0.006) and CDK4 (OR=1.14, p=0.04) with PCa risk. Machine learning (LASSO/SVM-RFE/Boruta) consistently selected ABCC4 and CDK4 as robust biomarkers, validated across independent cohorts. Single-cell analysis of 90,140 cells revealed 15 cell types, with significant expansion of myeloid dendritic cells (mDCs) in tumors. Both biomarkers showed dynamic expression along mDCs differentiation trajectories. Gene enrichment implicated focal adhesion, xenobiotic metabolism, and cell cycle pathways. Molecular docking predicted favorable binding poses for NP and OP with ABCC4 and CDK4, with binding energies below -4.5 kcal/mol. In vitro experiments confirmed that NP promoted PCa cell migration, while OP had no significant effect on cell proliferation. ABCC4 and CDK4 may serve as candidate mediators connecting NP/OP exposure with PCa. The expansion and metabolic reprogramming of mDCs suggest that mDCs may participate in tumor microenvironment remodeling in NP/OP-associated PCa. These findings provide molecular insights for environmental PCa prevention and highlight intervention targets.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:72db17d6b216f772d743bfe7716affd5e1c4dd30","kind":"journals","source":"Forests","title":"Odor Cues: Dongbei Bees for Olfactory Preference of T. amurensis Flower Volatiles","url":"https://doi.org/10.3390/f17111371","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.3390%2Ff17111371","date":"2026-11-17T00:00:00Z","timestamp":1794873600,"categories":["Docking & Screening"],"topic_ids":["docking"],"keywords":["Molecular docking"],"matched_keywords":["Molecular docking","compounds","chemical","Molecular"],"matched_tags":["docking","generic_chem"],"doi":"10.3390/f17111371","external_id":"72db17d6b216f772d743bfe7716affd5e1c4dd30","pdf_url":null,"code_url":null,"code_host":null,"authors":["Lan-Qing Li","Kai-Jia Xue","Hao-Ran Li","Lin Luo","Meng-Fei Tian","Quan-Yu Lin","Yu-Jiao Song","Chun-Ying Li","Chun-Jian Zhao"],"journal":"Forests","publisher":null,"impact_factor":null,"abstract":"Plant volatile organic compounds (VOCs) function as critical chemical signals mediating plant–pollinator interactions. Tilia amurensis Rupr. is an important protected wild plant species in China and a key nectar resource sustaining apiculture in Northeast China. The Dongbei bee (Apis mellifera ssp.), a nationally protected local honeybee breed, is an important pollinator of T. amurensis. Understanding the chemical communication between T. amurensis and the Dongbei bee is therefore essential for elucidating plant–pollinator interactions and promoting the conservation of forest biodiversity, honeybee genetic resources, and sustainable apiculture. In this study, using dynamic headspace sampling and GC–MS, we identified 32 flower VOCs emitted by T. amurensis flowers, with aldehydes identified as pivotal mediators. EAG was used to screen compounds that evoked robust antennal responses. Subsequent behavioral assays revealed that Dongbei bees were significantly attracted to heptanal at 10 µg/µL. Bee’s antennal transcriptome analysis after volatile stimulation revealed four candidate chemosensory genes (CSP3, CSP6, OBP21, and OR63) with high expression levels. Molecular docking analyses further suggested potential binding interactions between heptanal and the encoded proteins. Together, these findings suggest that heptanal serves as a key odor cue mediating the olfactory association between Dongbei bees and T. amurensis.","source_metadata":{"source":"semantic_scholar"}},{"id":"preprints:10.26434/chemrxiv.15012326/v1","kind":"preprints","source":"ChemRxiv","title":"PanJAK: predicting selectivity across the JAK family with deep learning","url":"https://doi.org/10.26434/chemrxiv.15012326/v1","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.26434%2Fchemrxiv.15012326%2Fv1","date":"2026-11-17T00:00:00Z","timestamp":1794873600,"categories":[],"topic_ids":[],"keywords":["molecular representations","kinase"],"matched_keywords":["deep learning","molecular representations","drug","molecular","ligand","compounds","chemical","compound","protein","kinase"],"matched_tags":["machine_learning","molecular_descriptor","generic_chem","biology"],"doi":"10.26434/chemrxiv.15012326/v1","external_id":"10.26434/chemrxiv.15012326/v1","pdf_url":null,"code_url":null,"code_host":null,"authors":["Sébastien Sueron","Francesca Grisoni"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"Predicting selectivity between closely related protein targets remains a major challenge in drug discovery. While drug-target affinity (DTA) models typically focus on predicting activity against individual targets, selectivity depends on resolving differences in activity between them. Here, we introduce panJAK, a DTA framework to predict selectivity across the four Janus kinase (JAK) paralogs. The model combines molecular representations with a ligand-conditioned encoding of structurally aligned binding pockets and predicts cross-paralog selectivity from differences in potency without requiring a selectivity-specific training objective. Using a curated dataset of approximately 10,000 compounds, we demonstrate that panJAK improves the quantitative prediction of selectivitywindows compared to established baseline models under scaffold-disjoint evaluation. Errors in individual potency predictions were positively correlated across paralogs and partly cancelled upon subtraction, resulting in more accurate selectivity estimates than expected under independent errors. Although prediction errors increased with chemical novelty, calibrated prediction intervals allowed for the identification of a subset of comparisons with high retrospective directional accuracy. These findings highlight the potential of adapting affinity-prediction models to address quantitative selectivity across homologous targets, supporting compound prioritization based on predicted selectivity profiles and associated reliability of those predictions.","source_metadata":{"date_field":"posted","keywords":[]}},{"id":"journals:10.1021/acsomega.6c11032","kind":"journals","source":"ACS Omega","title":"Penijavanicins A and B, Complex Polycyclic Polyketides\nFeaturing a\n                    cis\n                    -Fused Dihydroisocoumarin-Tetrahydropyran\nCore with Anti-Neuroinflammatory Activity","url":"https://doi.org/10.1021/acsomega.6c11032","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1021%2Facsomega.6c11032","date":"2026-11-17T00:00:00+00:00","timestamp":1794873600,"categories":["Docking & Screening"],"topic_ids":["docking"],"keywords":["molecular docking"],"matched_keywords":["molecular docking","Chemical","Compounds","compound","molecular"],"matched_tags":["docking","generic_chem"],"doi":"10.1021/acsomega.6c11032","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Min Hee Kim","Son Hung Tran","Gil Soo Kim","Ji-Eun Lee","Taehoon Oh","Gwi Ja Hwang","Hyung-Joo Jin","Sung-Kyun Ko","Young-Soo Hong","Bang Yeon Hwang","Suk Woo Kang","Jun-Pil Jang","Kyungsu Kang","Jae-Hyuk Jang"],"journal":"ACS Omega","publisher":"ACS","impact_factor":null,"abstract":"Chemical investigation of the soil-derived fungus Penicillium sp. 16F038 using the One Strain Many Compounds (OSMAC) strategy led to the discovery of penijavanicins A and B (1 and 2), two structurally novel aromatic polyketides. Comprehensive spectroscopic analyses, including 1,1-ADEQUATE NMR and HRESIMS, established their planar structures as an unprecedented 6/6/6/6/6 pentacyclic framework incorporating a cis-fused dihydroisocoumarin-tetrahydropyran architecture. The challenging stereochemical assignments, posed by the densely functionalized and conformationally constrained core, were successfully resolved through O-methylation derivatization, quantitative ROESY analysis, and TDDFT-ECD calculations. Biological evaluations revealed that compound 2 displayed significant antineuroinflammatory activity in LPS-stimulated BV-2 microglial cells by markedly suppressing nitric oxide (NO), TNF-α and IL-6 production without appreciable cytotoxicity. Furthermore, network pharmacology, molecular docking, and Western blot analyses suggested that 2 exerts its anti-inflammatory effects by inhibiting the ERK/NF-κB signaling pathway.","source_metadata":{"short_name":"ACS Omega","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"journals:8c20e19e277a93970a8550c396f69f8e44892900","kind":"journals","source":"Biophysica","title":"Pharmaceutical Drug Stability Analysis with Raman Spectroscopy for Spectral Degradation Assessment Using Deep Learning","url":"https://doi.org/10.3390/biophysica6060117","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.3390%2Fbiophysica6060117","date":"2026-11-17T00:00:00Z","timestamp":1794873600,"categories":["Spectra & Analytical"],"topic_ids":["chemometrics"],"keywords":["Convolutional Neural","CNN","Random Forest","XGBoost"],"matched_keywords":["Raman Spectroscopy for Spectral Degradation Assessment Using Deep Learning","Convolutional Neural","CNN","Deep Learning","Random Forest","XGBoost","Drug"],"matched_tags":["analytical_chemistry_ai","neural_network","machine_learning","ensemble","generic_chem"],"doi":"10.3390/biophysica6060117","external_id":"8c20e19e277a93970a8550c396f69f8e44892900","pdf_url":null,"code_url":null,"code_host":null,"authors":["V. A. Judi","T. Sasikala"],"journal":"Biophysica","publisher":null,"impact_factor":null,"abstract":"Background: Drug stability is a fundamental issue in drug-quality monitoring during storage. Methods: This study proposes a computational framework for degradation and potency estimation of pharmaceuticals using Raman spectroscopy. The absence of complete longitudinal Raman degradation data with experimental potencies led to the use of real Raman spectra as baseline spectra, which were then converted using physics-informed spectral degradation to create four stages (T0–T3). A modified 1D-Convolutional Neural Network (1D-CNN) comprising multi-scale convolution, residual learning, and spectral attention was designed to simultaneously classify the degradation stage and predict the remaining Active Pharmaceutical Ingredient potency. External computational testing of an independent set of commercial pharmaceutical dosage forms was performed using Raman data. The predicted potency profiles were further analyzed with an approximate zero-order kinetic model to estimate the degradation rate constant and the computational shelf-life t90. Results: The proposed modified 1D-CNN achieved better performance and an R2 value of 99.23 ± 0.02 for potency prediction, outperforming Support Vector Machine, Random Forest, and XGBoost models. From kinetic analysis, the R2 value was 0.98, the degradation rate constant was 5.78, and the calculated shelf-life was 1.76 normalized time units. Conclusions: The proposed framework can thus be considered as a computational proof-of-concept for Raman-based pharmaceutical degradation and potency assessment.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:4f5d1215dc0817f815a678d9de7126653d9bd4dc","kind":"journals","source":"Food Quality and Safety","title":"Photoprotective potential of tamarind (\n Tamarindus indica\n L.) seed coat polyphenols: extraction optimization, chemical profiling, and integrative computational analysis","url":"https://doi.org/10.1093/fqsafe/fyag088","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1093%2Ffqsafe%2Ffyag088","date":"2026-11-17T00:00:00Z","timestamp":1794873600,"categories":["Docking & Screening"],"topic_ids":["docking"],"keywords":["molecular docking","bioactivity"],"matched_keywords":["molecular docking","chemical","compounds","molecular","bioactivity"],"matched_tags":["docking","generic_chem","bioactivity"],"doi":"10.1093/fqsafe/fyag088","external_id":"4f5d1215dc0817f815a678d9de7126653d9bd4dc","pdf_url":null,"code_url":null,"code_host":null,"authors":["Yan-Mei Liu","Zi-Bo Song","Tao Wang","Chuan-Qi Chu","Yong-Li Jiang","Si-Yan Huang","Chun-Mei Yuan","Jun-Jie Yi"],"journal":"Food Quality and Safety","publisher":null,"impact_factor":null,"abstract":"Tamarind (Tamarindus indica L.) seed coat is an underutilized by-product of tamarind processing and is rich in polyphenols with potential antioxidant and photoprotective properties. However, the effects of extraction conditions on its phenolic composition and bioactivity remain unclear. This study investigated how solvent polarity and physical-assisted extraction technologies jointly influence the phenolic composition and explored the potential of tamarind seed coat polyphenols (TSCP) against ultraviolet (UV)-induced damage. Twenty-four combinations of four solvents and six extraction methods were systematically evaluated. Among the tested conditions, 75% ethanol combined with thermo-ultrasonic-assisted extraction showed high recovery of phenolic and flavonoid compounds, with total phenolic content (TPC, gallic acid equivalent (GAE)) and total flavonoid content (TFC, rutin equivalent (RE)) reaching 58.10 mg GAE/g extract and 224.36 mg RE/g extract, respectively. Ultra-high resolution liquid chromatography-mass spectrometry (UHR LC-MS) identified 40 phenolic compounds, mainly phenolic acids and flavonoids. The optimized extract exhibited strong radical-scavenging activity, with DPPH and ABTS radical-scavenging rates exceeding 98%, comparable to ascorbic acid. Network pharmacology and molecular docking further predicted potential targets and pathways associated with UV-induced damage, mainly involving DNA damage response, apoptosis, and the p53 signaling pathway. Overall, these findings demonstrate that extraction conditions substantially affect phenolic composition and support the potential of TSCP for UV-related protective applications, providing a promising strategy for the sustainable and value-added utilization of tamarind processing by-products.","source_metadata":{"source":"semantic_scholar"}},{"id":"preprints:10.26434/chemrxiv.15002429/v4","kind":"preprints","source":"ChemRxiv","title":"Pre-training on Quantum Mechanical Properties Improves ADMET Prediction","url":"https://doi.org/10.26434/chemrxiv.15002429/v4","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.26434%2Fchemrxiv.15002429%2Fv4","date":"2026-11-17T00:00:00Z","timestamp":1794873600,"categories":["ADMET & Safety"],"topic_ids":["admet"],"keywords":["ADMET Prediction","ADMET"],"matched_keywords":["ADMET Prediction","ADMET"],"matched_tags":["admet_prediction","admet"],"doi":"10.26434/chemrxiv.15002429/v4","external_id":"10.26434/chemrxiv.15002429/v4","pdf_url":null,"code_url":null,"code_host":null,"authors":["Arun Raja","Hongtao Zhao","Christian Tyrchan","Eva Nittinger","Michael M. Bronstein","Charlotte M. Deane","Garrett M. Morris"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"","source_metadata":{"date_field":"posted","keywords":[]}},{"id":"preprints:10.26434/chemrxiv.15012408/v1","kind":"preprints","source":"ChemRxiv","title":"Predicting Solvent-Induced Changes in Fluorescence Quantum Yield: Direct ClassificationVersus Regression","url":"https://doi.org/10.26434/chemrxiv.15012408/v1","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.26434%2Fchemrxiv.15012408%2Fv1","date":"2026-11-17T00:00:00Z","timestamp":1794873600,"categories":[],"topic_ids":[],"keywords":["Random Forest"],"matched_keywords":["Random Forest","molecule","molecules","compounds"],"matched_tags":["ensemble","generic_chem"],"doi":"10.26434/chemrxiv.15012408/v1","external_id":"10.26434/chemrxiv.15012408/v1","pdf_url":null,"code_url":null,"code_host":null,"authors":["Neethu Anand","Jibin Joseph"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"Changing the solvent can alter a molecule’s fluorescence quantum yield. This study compares two structure-based methods for predicting such changes: direct classification and prediction of two quantum yields followed by subtraction. From FluoDB-Lite, we retained 12,157 measurements for 2,712 molecules, resulting in 29,550 within-paper solvent comparisons. We categorized changes as reduced, approximately unchanged, or enhanced using a provisional absolute tolerance of 0.10. Molecules were divided into training, validation, and test sets without identity overlap. Validation identified a histogram gradient-boosting classifier and a Random Forest regressor as optimal models. On 407 test molecules, molecule-weighted macro-F1 scores were 0.577 and 0.569, respectively, compared with 0.248 for the majority baseline. The paired bootstrap interval for their difference included zero. The mean absolute error for individual quantum-yield prediction was 0.185. Small external checks and a design screen demonstrate the use and limitations of these approaches. Both methods capture useful information, but neither reliably improves experimental results for new compounds.","source_metadata":{"date_field":"posted","keywords":[]}},{"id":"preprints:10.26434/chemrxiv.15012395/v1","kind":"preprints","source":"ChemRxiv","title":"Predicting the Host-guest Binding Gibbs Free Energy for Guests","url":"https://doi.org/10.26434/chemrxiv.15012395/v1","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.26434%2Fchemrxiv.15012395%2Fv1","date":"2026-11-17T00:00:00Z","timestamp":1794873600,"categories":["Cheminformatics"],"topic_ids":["cheminformatics"],"keywords":["molecular fingerprints","Gibbs Free Energy"],"matched_keywords":["molecular fingerprints","Gibbs Free Energy","molecules","drug","molecular"],"matched_tags":["molecular_fingerprints","thermodynamics","generic_chem"],"doi":"10.26434/chemrxiv.15012395/v1","external_id":"10.26434/chemrxiv.15012395/v1","pdf_url":null,"code_url":null,"code_host":null,"authors":["Rauno Reitalu","Riina Aav","Mario Öeren"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"Host-guest chemistry involves noncovalent interactions between host and guest molecules and has been applied in areas such as adaptive materials, sensing, and drug delivery. However, the discovery of new hosts and the identification of suitable guests remain challenging because selective binding is governed by subtle and often competing noncovalent interactions that are difficult to predict. To overcome this predictive limitation, computational framework was developed to predict Gibbs free energies of host-guest binding using a hybrid machine-learning model. The model was trained and tested on 4155 host-guest systems manually curated from the literature and combines descriptors derived from semiempirical quantum-mechanical calculations with three-dimensional molecular fingerprints. Evaluation on a test set yielded an R2 of 0.83 and a mean absolute error of 3.1 kJ per mol. The computational efficiency of the model enabled prediction times of approximately one to five minutes per host-guest system with an average size of 174 atoms, supporting the framework’s applicability to high-throughput screening for host-guest discovery.","source_metadata":{"date_field":"posted","keywords":[]}},{"id":"preprints:10.26434/chemrxiv.15012400/v1","kind":"preprints","source":"ChemRxiv","title":"Protein chemical shift prediction in web browsers: seven major programs ported, benchmarked, and studied for what they help to reveal about NMR structures, AlphaFold models and molecular simulations","url":"https://doi.org/10.26434/chemrxiv.15012400/v1","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.26434%2Fchemrxiv.15012400%2Fv1","date":"2026-11-17T00:00:00Z","timestamp":1794873600,"categories":[],"topic_ids":[],"keywords":["AlphaFold 3","Claude","molecular dynamics","MD simulations"],"matched_keywords":["AlphaFold 3","neural network","Claude","chemical","molecular","molecular dynamics","MD simulations","Protein"],"matched_tags":["ml_protein_ligand","neural_network","LLMs","generic_chem","molecular_dynamics","biology"],"doi":"10.26434/chemrxiv.15012400/v1","external_id":"10.26434/chemrxiv.15012400/v1","pdf_url":null,"code_url":null,"code_host":null,"authors":["Luciano A. Abriata"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"Motivation. Predicting NMR chemical shifts from protein structures is routine in structure validation, resonance assignment and the analysis of molecular dynamics simulations, but the available programs come as platform-specific executables, aging software stacks or upload-based servers, and some are no longer distributed. Results. Assisted by Claude Code, I ported seven chemical shift predictors to run entirely in the web browser with no installation and no data leaving the user's computer: UCBShift-X, SHIFTX2, SPARTA+, PPM_One (with its MD-ensemble variant PPM), CamShift, LEGOLAS and the random-coil predictor POTENCI. The web apps share similar interfaces that work for single structures or ensembles (from NMR, MD simulations and emulations, docking poses, etc.), and each reproduces its original implementation to printed precision. On 153 proteins with BMRB assignments, UCBShift-X is the most accurate structure-based predictor (best for 71 % of proteins), followed by SHIFTX2, SPARTA+ and PPM_One, and then by LEGOLAS and CamShift. However, I found that the input structure matters more than the predictor: For every program, AlphaFold models and crystal structures give about 15 % lower errors than the deposited NMR structures! Averaging over MD trajectories also improves most programs, especially CamShift, LEGOLAS and PPM. The ranking holds on three recently determined NMR structures not yet released at the PDB/BMRB, and comparing predictions on the NMR and AlphaFold 3 models shows which parts of each structure the shifts support and which parts they flag as potential outliers or assignment errors. Finally, I ported SLAE-CS, a very recent neural network that predicts only backbone 15N shifts: on the deposited NMR structures it is about as accurate as UCBShift-X, but on AlphaFold models and crystal structures its 15N errors are 17–28 % lower. Availability. Static web applications at https://pdbms.altervista.org/web-csp/ within the PDB Manipulation Suite platform (https://pdbms.altervista.org). For a full report with all methods, tables and data, see the semi-automated, extended version of this preprint at https://pdbms.altervista.org/web-csp/predictions_report/.","source_metadata":{"date_field":"posted","keywords":[]}},{"id":"journals:8b8bf6889b478d191ca9950cfdc0a181326a9762","kind":"journals","source":"Food & function","title":"Protein language model-assisted screening of bitter peptides from Prunus dulcis","url":"https://doi.org/10.1039/d6fo02313d","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1039%2Fd6fo02313d","date":"2026-11-17T00:00:00Z","timestamp":1794873600,"categories":["Docking & Screening","Targets & Structures"],"topic_ids":["docking","targets"],"keywords":["Molecular docking","chemicals","receptor"],"matched_keywords":["Molecular docking","machine learning","chemicals","Molecular","ligand","Protein","receptor"],"matched_tags":["docking","machine_learning","generic_chem","biology"],"doi":"10.1039/d6fo02313d","external_id":"8b8bf6889b478d191ca9950cfdc0a181326a9762","pdf_url":null,"code_url":null,"code_host":null,"authors":["Yu-Yang Liu","Sai-Nan Yu","Yan Zhang","Jing-Xian Zheng","Xue-Qi Fu","Yue Jin"],"journal":"Food & function","publisher":null,"impact_factor":null,"abstract":"Bitterness is a major limitation in the utilization of plant-derived proteins in food systems, while conventional identification of bitter peptides is labor-intensive and inefficient. Advances in food informatics, particularly protein language models, provide new opportunities for predicting the properties of food chemicals. In this study, we developed a protein language model-assisted workflow to screen and prioritize potential bitter peptides from Prunus dulcis proteins. Protein sequences were first subjected to in silico digestion to generate candidate peptides. Representations derived from ESM Cambrian (ESMC) embeddings were combined with amino acid index (AAI) descriptors and Binary Profile-based Numerical Coding (BPNC) features to construct a machine learning model for bitterness prediction. The resulting candidates were further prioritized based on their interaction with the bitter taste receptor TAS2R14 using the ConPLex framework. Three peptides, GECIL, GRPR, and GIPEV, were selected for further evaluation. HPLC analysis revealed hydrolysate peaks with retention times consistent with those of the corresponding synthetic standards, providing preliminary chromatographic evidence, and electronic tongue measurements indicated bitterness-related instrumental responses. CCK-8 assays indicated generally maintained cell viability, except for a decrease with GIPEV at 3.2 mg mL-1. Molecular docking and dynamics simulations suggested stable interaction profiles with TAS2R14, with residues S265 and Q266 likely involved in ligand recognition. This study presents an integrated food informatics workflow combining protein language models with structural and experimental evaluation. The proposed strategy provides a useful approach for discovering bitterness-related peptides from plant proteins and offers insight into the molecular basis of peptide bitterness.","source_metadata":{"source":"semantic_scholar"}},{"id":"preprints:10.26434/chemrxiv.15000193/v2","kind":"preprints","source":"ChemRxiv","title":"Reaction-Dependent Predictability of Yields from Small Substrate-Scope Datasets","url":"https://doi.org/10.26434/chemrxiv.15000193/v2","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.26434%2Fchemrxiv.15000193%2Fv2","date":"2026-11-17T00:00:00Z","timestamp":1794873600,"categories":["Cheminformatics"],"topic_ids":["cheminformatics"],"keywords":["molecular fingerprints"],"matched_keywords":["molecular fingerprints","molecular"],"matched_tags":["molecular_fingerprints","generic_chem"],"doi":"10.26434/chemrxiv.15000193/v2","external_id":"10.26434/chemrxiv.15000193/v2","pdf_url":null,"code_url":null,"code_host":null,"authors":["Shinya shiomi","Takuma Morimoto","Masahiro Yoshida"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"We evaluate structure-based regression using molecular fingerprints and ElasticNet across four reaction datasets (96–190 reactions). The models outperform constant-yield baselines for two transformations, including literature-assembled umpolung amide synthesis. Learning curves reveal reaction-dependent data requirements, with useful predictions emerging from 20 training examples in favorable cases.","source_metadata":{"date_field":"posted","keywords":[]}},{"id":"journals:66affec8977974dac9af76232c8c0e9653a850a6","kind":"journals","source":"Pharmacological reports","title":"Selected carboxyl-group bioisosteres of mycophenolic acid with divergent biological profiles","url":"https://doi.org/10.1007/s43440-026-00931-w","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1007%2Fs43440-026-00931-w","date":"2026-11-17T00:00:00Z","timestamp":1794873600,"categories":["Docking & Screening"],"topic_ids":["docking"],"keywords":["molecular docking","enzyme","bioisosteres","Bioisosteric replacement"],"matched_keywords":["molecular docking","enzyme","compounds","molecular","bioisosteres","Bioisosteric replacement"],"matched_tags":["docking","biology","generic_chem","bioisosteres"],"doi":"10.1007/s43440-026-00931-w","external_id":"66affec8977974dac9af76232c8c0e9653a850a6","pdf_url":null,"code_url":null,"code_host":null,"authors":["J. Walczak","D. Iwaszkiewicz-Grześ","D. Garbicz","Alan Bielecki","L. Persoons","D. Schols","Saba Shahrivari-Baviloliaei","S. de Jonghe","Agnieszka Viapiana","A. Plenis","Mario Smet","G. Cholewiński"],"journal":"Pharmacological reports : PR","publisher":null,"impact_factor":null,"abstract":"BACKGROUND Mycophenolic acid (MPA) is an immunosuppressive inhibitor of inosine-5'-monophosphate dehydrogenase (IMPDH). Bioisosteric replacement of its carboxyl group may modulate potency, cellular selectivity, and the overall biological profile. This study compared cyclic 1,2,4-oxadiazole derivatives and non-cyclic carboxylate bioisosteres to assess their effects on histone deacetylase 6 (HDAC6), IMPDH-related activity, redox properties, and cellular responses. METHODS The compounds were synthesised by dehydrative cyclisation of MPA with amidoximes or by coupling reactions affording hydroxamic acid, hydrazide, and amide analogues. They were evaluated by molecular docking to IMPDH type 2 (IMPDH2) and HDAC6, HDAC6 enzymatic inhibition, metabolic cytotoxicity and flow-cytometric proliferation assays in peripheral blood mononuclear cells (PBMCs) and T-Jurkat cells, guanosine monophosphate (GMP) rescue, and 2,2-diphenyl-1-picrylhydrazyl radical scavenging. RESULTS Docking showed scaffold-dependent binding patterns. Non-cyclic bioisosteres showed more MPA-like predicted IMPDH2 binding modes and, when bearing zinc-binding groups, were more compatible with the zinc-containing HDAC6 catalytic region, whereas oxadiazoles adopted alternative predicted binding modes. In the HDAC6 enzymatic assay, pyridyl oxadiazoles were the most active cyclic derivatives, while the hydroxamic acid derivative showed nanomolar inhibition. In T-Jurkat cells, the newly synthesised derivatives showed substantially lower cytotoxic potency than MPA and mycophenolate mofetil (MMF), while several retained measurable antiproliferative activity. GMP rescue suggested functional divergence: non-cyclic bioisosteres showed GMP-rescuable antiproliferative effects consistent with guanine nucleotide depletion, whereas selected oxadiazoles were not rescued under these conditions. The antioxidant assay highlighted hydroxamic and hydrazide derivatives and the 2-furanyl oxadiazole. CONCLUSIONS Carboxyl-group bioisosterism of MPA reshaped enzyme inhibition, redox activity, and cellular profiles. Non-cyclic derivatives retained GMP-rescuable antiproliferative profiles, with the hydroxamic acid analogue exhibiting particularly strong HDAC6 inhibition, whereas 1,2,4-oxadiazoles generally showed reduced cytotoxic potency and distinct cellular response patterns.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:10.1186/s13321-026-01333-z","kind":"journals","source":"Journal of Cheminformatics","title":"SKiP: a Structure-oriented Kinetics Prediction suite for enzyme-substrate complexes","url":"https://doi.org/10.1186/s13321-026-01333-z","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1186%2Fs13321-026-01333-z","date":"2026-11-17T00:00:00+00:00","timestamp":1794873600,"categories":["Cheminformatics"],"topic_ids":["cheminformatics"],"keywords":[],"matched_keywords":[],"matched_tags":["biology"],"doi":"10.1186/s13321-026-01333-z","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Sowmya Ramaswamy Krishnan","Nishtha Pandey","Arijit Roy"],"journal":"Journal of Cheminformatics","publisher":"Springer","impact_factor":5.7,"abstract":"","source_metadata":{"short_name":"J. Cheminform.","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"preprints:10.26434/chemrxiv.15012327/v1","kind":"preprints","source":"ChemRxiv","title":"Stable Pt Single Atoms Embedded in Ge(001) Enable Efficient Hydrogen Evolution","url":"https://doi.org/10.26434/chemrxiv.15012327/v1","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.26434%2Fchemrxiv.15012327%2Fv1","date":"2026-11-17T00:00:00Z","timestamp":1794873600,"categories":[],"topic_ids":[],"keywords":["density functional theory"],"matched_keywords":["ML","density functional theory"],"matched_tags":["machine_learning","dft"],"doi":"10.26434/chemrxiv.15012327/v1","external_id":"10.26434/chemrxiv.15012327/v1","pdf_url":null,"code_url":null,"code_host":null,"authors":["Idoia Camara","Andoni Ugartemendia","Ramon M. Bergua","Jose M. Mercero","Elisa Jimenez-Izal"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"Single-atom catalysts offer a promising route to maximize noble-metal utilization, but maintaining isolated active sites under operating conditions remains a major challenge. Pt deposition on Ge(001) induces the spontaneous formation of thermodynamically stable nanowire arrays in which Pt atoms remain atomically dispersed within the Ge lattice, creating well-defined single-atom catalytic sites. Here, we investigate the hydrogen evolution reaction (HER) on Pt–Ge(001) using grand-canonical density functional theory under constant-potential conditions, explicitly accounting for the effects of electrode potential, hydrogen coverage, solvent effects, and pH. Hydrogen adsorbs selectively on Pt sites, while the Ge environment stabilizes the Pt atoms and tunes their hydrogen-binding properties. The calculated activation barriers favor a Volmer–Heyrovsky pathway, with facile Volmer adsorption promoting hydrogen accumulation under cathodic conditions. At 1 ML hydrogen coverage, the calculated Volmer and Heyrovsky barriers are substantially lower than reported values for Pt(111), indicating competitive HER kinetics. These results identify Pt–Ge(001) nanowire arrays as a promising platform for the HER and demonstrate how the interplay between support-induced stabilization, electrode potential, and adsorbate coverage can govern the activity and reaction mechanism of single-atom electrocatalysts.","source_metadata":{"date_field":"posted","keywords":[]}},{"id":"journals:a7a01cb59ac2a813965b25a30721fb4a4f874911","kind":"journals","source":"Molecular & cellular proteomics","title":"Structural and Functional Proteomics Reveal the Molecular Response of Human Lung Cells to Paraoxon Exposure","url":"https://doi.org/10.1016/j.mcpro.2026.101705","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1016%2Fj.mcpro.2026.101705","date":"2026-11-17T00:00:00Z","timestamp":1794873600,"categories":["Docking & Screening"],"topic_ids":["docking"],"keywords":["molecular docking","chemicals","enzyme"],"matched_keywords":["molecular docking","Molecular","chemicals","compound","chemical","protein","enzyme"],"matched_tags":["docking","generic_chem","biology"],"doi":"10.1016/j.mcpro.2026.101705","external_id":"a7a01cb59ac2a813965b25a30721fb4a4f874911","pdf_url":null,"code_url":null,"code_host":null,"authors":["Tong Zhang","Snigdha Sarkar","Victoria N. Prozapas","Chathuri J. Kombala","A. Shouaib","M. Berger","Chelsea M. Hutchinson-Bunch","W. Chrisler","Kristin M. Engbrecht","J. Eder","J. Kyle","I. Attah","Rosalie K. Chu","Song-Jiang Feng","Ivo Díaz Ludovico","Vivian S. Lin","J. Melchior"],"journal":"Molecular & cellular proteomics : MCP","publisher":null,"impact_factor":null,"abstract":"Technologies that facilitate rapid and comprehensive assessment of emerging chemicals of concern will enable toxicologists to respond promptly and effectively to mitigate their harmful effects. In this study, we demonstrate the utility of integrating advanced proteomics techniques including activity-based protein profiling (ABPP), thermal proteome profiling (TPP), and limited proteolysis-coupled mass spectrometry (LiP-MS) in uncovering how xenobiotic agents influence biological pathways. Using paraoxon, a well-studied organophosphorus (OP) compound, we found that treatment of human lung epithelial cells inhibited enzymes involved in lipid metabolism and induced corresponding changes in the lipidome. ABPP revealed selective inhibition of hydrolases by paraoxon, while TPP uncovered dose-dependent thermal stability changes in specific proteins, including stabilization of protein phosphatase methylesterase (PPME1) and destabilization of its substrate protein phosphatase 2A (PP2A). Using LiP-MS combined with molecular docking, we identified multiple putative paraoxon binding sites on the large, multifunctional enzyme fatty acid synthase (FASN), including unexpected interactions at the ketoacyl synthase and dehydratase domains. These paraoxon-induced perturbations suggest cellular toxicity may be related to the disruption of critical biomolecular interactions orchestrating G2/M transition, leading to delayed entry into mitosis, and the regulation on lipid metabolism. This work demonstrates how integration of chemoproteomic and structural proteomic techniques can provide complementary insights into mechanisms contributing to chemical toxicity.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:baf984e9badee55e42ec05e75a926eb43f58f24c","kind":"journals","source":"PloS one","title":"Structure-based computational screening of cyclotides as putative CYP19A1-binding scaffolds for modulation of estrogen biosynthesis in adenomyosis","url":"https://doi.org/10.1371/journal.pone.0352405","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1371%2Fjournal.pone.0352405","date":"2026-11-17T00:00:00Z","timestamp":1794873600,"categories":[],"topic_ids":[],"keywords":["computational screening","cytochrome P450","molecular dynamics","receptor"],"matched_keywords":["computational screening","molecular","cytochrome P450","molecular dynamics","receptor","protein"],"matched_tags":["virtual_screening","generic_chem","toxicity","molecular_dynamics","biology"],"doi":"10.1371/journal.pone.0352405","external_id":"baf984e9badee55e42ec05e75a926eb43f58f24c","pdf_url":null,"code_url":null,"code_host":null,"authors":["S. Üstünel","Nouman Ali"],"journal":"PloS one","publisher":null,"impact_factor":null,"abstract":"Adenomyosis is a chronic estrogen-dependent gynecological disorder associated with dysmenorrhea, abnormal uterine bleeding, chronic pelvic pain, infertility, and recurrent implantation failure. Because human placental aromatase cytochrome P450 (CYP19A1) plays a central role in local estrogen biosynthesis, this study employed an integrated in silico workflow to identify cyclotide scaffolds capable of interacting with CYP19A1. The receptor structure was prepared and validated, while protein-protein interaction analysis confirmed its involvement in steroidogenesis and estrogen biosynthesis. Twenty cyclotides were modeled and screened against the predicted active site of CYP19A1. Based on an integrated evaluation of HADDOCK and ClusPro docking, binding orientation, and interaction characteristics, cycloviolacin O4 was prioritized as the lead candidate. The predicted CYP19A1-cycloviolacin O4 complex exhibited a HADDOCK score of -82.8 ± 9.8, a cluster RMSD of 0.2 ± 0.1 Å, and a ClusPro energy score of -872.1. Docking against the reference inhibitor exemestane and the reference peptide SNA-60-367-10 provided additional computational benchmarking. Subsequent 500 ns molecular dynamics simulations, normal mode analysis, and MM-GBSA/MM-PBSA calculations supported the predicted structural stability and dynamic compatibility of the complex, yielding binding free energies of -66.58 ± 3.46 kcal/mol (MM-GBSA) and -4.56 ± 4.29 kcal/mol (MM-PBSA). Computational safety assessment predicted cycloviolacin O4 to be non-antigenic, non-allergenic, and non-toxic, while immune simulation suggested a balanced immune response. Collectively, these findings identify cycloviolacin O4 as a computationally predicted CYP19A1-binding cyclotide scaffold. However, these observations are derived exclusively from computational analyses and require biochemical, cellular, and in vivo validation before any conclusions regarding biological activity or therapeutic potential can be drawn.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:f99f524b28d0307c529bdbe0cc343eed57c0ef17","kind":"journals","source":"RSC advances","title":"Synthesis and antiproliferative screening of para-halogenated 4H-chromene-benzene hybrids: crystallographic, and molecular-docking studies","url":"https://doi.org/10.1039/d6ra08212b","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1039%2Fd6ra08212b","date":"2026-11-17T00:00:00Z","timestamp":1794873600,"categories":[],"topic_ids":[],"keywords":["Molecular docking","ADMET","pharmacokinetic","IC50","CYP inhibition","kinase"],"matched_keywords":["Molecular docking","ADMET","pharmacokinetic","IC50","CYP inhibition","molecular","compounds","Compound","kinase"],"matched_tags":["docking","admet","bioactivity","toxicity","generic_chem","biology"],"doi":"10.1039/d6ra08212b","external_id":"f99f524b28d0307c529bdbe0cc343eed57c0ef17","pdf_url":null,"code_url":null,"code_host":null,"authors":["M. Helal","Z. Eldin","S. Alrashdi","Z. Moussa","Walid E. Elgammal","A. Halawa","A. Elhenawy","A. M. el-Agrody"],"journal":"RSC advances","publisher":null,"impact_factor":null,"abstract":"A series of novel 4H-chromene-benzene hybrids bearing para-halogen substituents (F, Cl, Br, I) on the phenyl moiety were efficiently synthesized via an ultrasonic-assisted one-pot three-component cyclocondensation reaction. The reaction proceeded in 93-96% yield within 2 minutes under mild conditions. The structures were fully characterized by IR, 1D/2D NMR, mass spectrometry, and single-crystal X-ray diffraction of 4b. The antiproliferative activity of all compounds was evaluated against twelve human cancer cell lines using the MTT assay. Compound 4b (4-Cl) exhibited the most broadly active profile (mean IC50 = 16.4 µM), while 4c (4-Br) was the least active (mean IC50 = 29.6 µM). The SAR was found to be cell-line-dependent, with 4a (4-F) showing superior activity against MDA-MB-231, HCT-116, and SW480. Molecular docking against the EGFR kinase domain (PDB: 4HJO) revealed that 4d (4-I) achieved the most favorable binding energy (-6.953 kcal mol-1) and the lowest RMSD (1.878 Å), though these computational predictions require experimental validation. In silico ADMET profiling predicted high GI absorption and minimal CYP inhibition for 4a-c. Compound 4b emerged as the most promising lead based on its balanced activity and favorable pharmacokinetic profile. The proposed EGFR-mediated mechanism remains hypothetical and warrants experimental confirmation.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:127e573dd9b7134dad2d1a4845b42b599f57bccf","kind":"journals","source":"Molecules","title":"Tryptanthrin-6-Oxime Induces Apoptosis in SW620 Cells and Inhibits Tumor Growth in a Xenograft Mouse Model","url":"https://doi.org/10.3390/molecules31224170","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.3390%2Fmolecules31224170","date":"2026-11-17T00:00:00Z","timestamp":1794873600,"categories":["Docking & Screening"],"topic_ids":["docking"],"keywords":["pharmacophore screening","kinase"],"matched_keywords":["pharmacophore screening","compounds","compound","drug","kinase"],"matched_tags":["pharmacophore","generic_chem","biology"],"doi":"10.3390/molecules31224170","external_id":"127e573dd9b7134dad2d1a4845b42b599f57bccf","pdf_url":null,"code_url":null,"code_host":null,"authors":["E. Zavjalov","O. Kichakova","L. D. Romanenko","I. Razumov","M. Buldakov","Kristina D. Vasilenko","Ksenia S. Khludneva","I. Schepetkin","A. V. Uvarov","A. R. Kovrizhina","A. Khlebnikov"],"journal":"Molecules","publisher":null,"impact_factor":null,"abstract":"Colorectal cancer (CRC) remains a major clinical challenge, and the c-Jun N-terminal kinase (JNK) signaling pathway represents a compelling yet pharmacologically complex target. Two tetracyclic ketoxime JNK inhibitors, 11H-indeno[1,2-b]quinoxalin-11-one oxime (IQ-1) and 6-(hydroxyimino)indolo[2,1-b]quinazolin-12(6H)-one (tryptanthrin-6-oxime, TRP-Ox), have previously demonstrated anti-inflammatory activity, but their direct antitumor potential against CRC has not been systematically investigated. In this study, we evaluated the in vitro and in vivo efficacy of both compounds against human SW620 colon adenocarcinoma models. Reverse pharmacophore screening of the compound’s isomers confirmed JNK1 and JNK3 as primary predicted targets alongside shared cancer-relevant effectors (including CDK2, Cyclin A2, VEGFR2, GSK3β, PTP1B, HSP90α, and MMP3), suggesting a polypharmacological mode of action. In vitro proliferation was monitored continuously via real-time impedance analysis (iCELLigence), while cell death was quantified after 24 h using annexin V/propidium iodide flow cytometry. Both compounds suppressed SW620 cell growth, driving cytotoxicity predominantly through early and late apoptosis. In vivo, immunodeficient SCID mice bearing heterotopic SW620 xenografts were treated daily for 28 days with IQ-1 (30 mg/kg) or TRP-Ox (30, 60, or 90 mg/kg) via intraperitoneal (i.p.) or peritumoral (subcutaneous, s.c.) injection, with oxaliplatin (5 mg/kg, i.p.) as a positive control. Unlike oxaliplatin, IQ-1 showed no antitumor activity by either route. Conversely, TRP-Ox induced xenograft tissue necrosis and dose-dependently reduced post-mortem tumor mass at 30 and 60 mg/kg via both administration routes. However, the 90 mg/kg dose yielded no further reduction, indicating an efficacy plateau likely associated with dose-limiting toxicity. Notably, peritumoral administration significantly outperformed delivery, indicating that the systemic efficacy of TRP-Ox is constrained by low bioavailability. Collectively, these findings identify TRP-Ox as a potent pro-apoptotic lead active against CRC cells and warrant its further preclinical development using advanced drug-delivery formulations.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:4b7305c6efa5d02f61631580a872f4bd1eee632f","kind":"journals","source":"ASN neuro","title":"TUBB2A as a Candidate Therapeutic Target for Alzheimer's Disease: Orotidine-Mediated Metabolic Mechanisms and Druggability","url":"https://doi.org/10.1080/17590914.2026.2754640","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1080%2F17590914.2026.2754640","date":"2026-11-17T00:00:00Z","timestamp":1794873600,"categories":["Docking & Screening","Targets & Structures"],"topic_ids":["docking","targets"],"keywords":["molecular docking","binding affinity"],"matched_keywords":["molecular docking","molecular","binding affinity"],"matched_tags":["docking","generic_chem","bioactivity"],"doi":"10.1080/17590914.2026.2754640","external_id":"4b7305c6efa5d02f61631580a872f4bd1eee632f","pdf_url":null,"code_url":null,"code_host":null,"authors":["Li-Ling Liang","Jing-Jing Zhang","Ying Zhang"],"journal":"ASN neuro","publisher":null,"impact_factor":null,"abstract":"As the most prevalent neurodegenerative disorder, Alzheimer's disease (AD) poses a pressing need for novel therapeutic targets. This study was designed to address this gap by screening and validating druggable genetic targets nominally associated with AD through an integrative multi-omics approach. Mendelian randomization (MR) screened the druggable genome. GTEx eQTL data and differential expression analysis identified core candidates. SMR, GSMR, and cross-tissue eQTL validation assessed putative causal effects and cross-tissue generalizability. Single-cell RNA sequencing (scRNA-seq) dissected cell type-specific expression. Virtual knockout combined with cell-cell communication analysis elucidated regulatory mechanisms. Mediation analysis evaluated Cerebrospinal fluid (CSF) metabolite involvement, and molecular docking screened therapeutic agents. Genetically predicted TUBB2A expression was nominally associated with AD risk. Virtual knockout analysis revealed that TUBB2A deficiency induced significant upregulation of SLC1A3 and HIF3A, suggesting compensatory glutamate clearance and hypoxic stress responses. Nominal mediation analysis suggested that the CSF metabolite orotidine may be involved in the TUBB2A-AD association. Molecular docking revealed high binding affinity between Eribulin mesylate and TUBB2A. This study identifies TUBB2A as a priority candidate gene for further validation in AD, proposes a potential role for orotidine in TUBB2A-related AD susceptibility, and identifies eribulin mesylate as a potential therapeutic agent worthy of further investigation, thereby generating testable hypotheses for future mechanistic and therapeutic studies. susceptibility, and identifies eribulin mesylate as a potential therapeutic agent worthy of further investigation, thereby generating testable hypotheses for future mechanistic and therapeutic studies.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:10.1038/s41597-026-08302-z","kind":"journals","source":"Scientific Data","title":"Unifying biomedical knowledge in a modern multimodal graph","url":"https://doi.org/10.1038/s41597-026-08302-z","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1038%2Fs41597-026-08302-z","date":"2026-11-17T00:00:00+00:00","timestamp":1794873600,"categories":["LLMs & Agents"],"topic_ids":["llm"],"keywords":[],"matched_keywords":["machine learning","molecular"],"matched_tags":["machine_learning","generic_chem"],"doi":"10.1038/s41597-026-08302-z","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Lucas Vittor","Ayush Noori","Iñaki Arango","Joaquín Polonuer","Sam Rodriques","Andrew White","David A. Clifton","Marinka Zitnik"],"journal":"Scientific Data","publisher":"Nature","impact_factor":6.9,"abstract":"Biomedical knowledge graphs (KGs) are widely used in the life sciences, yet many are derived from unstructured documents and therefore lack schema-level constraints, whereas graphs assembled from structured resources are difficult to harmonize into a unified representation. We present O ptimus KG, a multimodal biomedical labeled property graph (LPG) built from structured and semi-structured resources to preserve factual, type-specific metadata across molecular, anatomical, clinical, and environmental domains. O ptimus KG contains 190,939 nodes across 10 entity types, 21,818,752 edges across 27 edge types, and 67,070,490 property instances encoding 109,665,797 values across 145 distinct property keys, derived from 18 ontologies and controlled vocabularies. The graph enforces a top-level schema for nodes and edges and retains granular, type-specific properties, cross-references, and provenance. We assessed the validity of O ptimus KG by evaluating whether graph relationships are supported by evidence from the scientific literature using a multimodal agent, PaperQA3. PaperQA3 identified supporting evidence for 70.0% of sampled edges, whereas 83.4% of sampled false edges received no supporting evidence. Edges without literature support were concentrated in associations derived from experimental and functional genomics resources, suggesting that O ptimus KG captures biomedical knowledge that may precede synthesis in the scientific literature. O ptimus KG is distributed as Apache Parquet files, providing a standardized resource for graph-based machine learning, knowledge-grounded retrieval with large language models, and biomedical discovery use cases such as hypothesis generation.","source_metadata":{"short_name":"Sci. Data","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"preprints:2611.21857v1","kind":"preprints","source":"arXiv","title":"tmol: a GPU-accelerated implementation of the Rosetta energy function for machine learning","url":"https://arxiv.org/abs/2611.21857v1","detail_url":"/chemradar/article?u=https%3A%2F%2Farxiv.org%2Fabs%2F2611.21857v1","date":"2026-11-16T20:20:33Z","timestamp":1794860433,"categories":[],"topic_ids":[],"keywords":[],"matched_keywords":["machine learning","ML","protein"],"matched_tags":["machine_learning","biology"],"doi":null,"external_id":"2611.21857v1","pdf_url":"https://arxiv.org/pdf/2611.21857v1","code_url":"https://github.com/RosettaCommons/tmol","code_host":"GitHub","authors":["Andrew Leaver-Fay","Kieran Didi","Jeff Flatten","Alex Ford","Guangfeng Zhou","Nathaniel Corley","Hahnbeom Park","Maxwell W. Kazman","Rafael I. Brent","Joseph Kleinhenz","Henry Solberg","Andrew M. Watkins","Brian Kuhlman","David Baker","Frank DiMaio"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"Physics-inspired forcefields complement generative machine learning models, showing better generalization beyond the training distribution. However, these forcefields are typically slow to evaluate, and their derivatives are difficult to integrate into ML frameworks. We introduce tmol, a GPU-native PyTorch library implementing the state-of-the-art RosettaGenFF forcefield. tmol supports efficient batched differentiable scoring and whole-structure refinement with up to 362-fold throughput increase (batched). Used as a reward in generative protein design, tmol increases hydrogen bonds and water bridges at designed interfaces.","source_metadata":{"categories":["cs.LG"],"code_status":"found","code_url":"https://github.com/RosettaCommons/tmol"}},{"id":"journals:10.1039/d6dd00428h","kind":"journals","source":"Digital Discovery","title":"Mapping structured absences in scientific papers: an auditable LLM-assisted workflow demonstrated on cross-coupling chemistry","url":"https://doi.org/10.1039/d6dd00428h","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1039%2Fd6dd00428h","date":"2026-11-16T19:43:16Z","timestamp":1794858196,"categories":["Reaction Informatics","LLMs & Agents"],"topic_ids":["reaction","llm"],"keywords":["LLM","cross coupling"],"matched_keywords":["LLM","cross-coupling"],"matched_tags":["LLMs","named_coupling_reaction"],"doi":"10.1039/d6dd00428h","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Yuriko Ono","Masaharu Yoshioka","Tetsuya Taketsugu"],"journal":"Digital Discovery","publisher":"RSC","impact_factor":5.6,"abstract":"Scientific papers record what was discovered and communicated, but the resolution at which evidence relations are made explicit — a paper’s documentary resolution — often falls short of what later...","source_metadata":{"short_name":"Digital Discovery","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"preprints:2611.21423v1","kind":"preprints","source":"arXiv","title":"Emergent Steering Beyond Endpoint Alignment in Chemical Reaction Models","url":"https://arxiv.org/abs/2611.21423v1","detail_url":"/chemradar/article?u=https%3A%2F%2Farxiv.org%2Fabs%2F2611.21423v1","date":"2026-11-16T17:13:17Z","timestamp":1794849197,"categories":["Reaction Informatics"],"topic_ids":["reaction"],"keywords":["reaction prediction","retrosynthesis"],"matched_keywords":["reaction prediction","model represents chemical transformation or why it fails. We introduce UniRxnAxis, a unified forward-retrosynthesis","retrosynthesis, it recovers most of the model","Machine learning","Chemical","retrosynthesis"],"matched_tags":["reaction_prediction","retrosynthesis_method","machine_learning","generic_chem","retrosynthesis"],"doi":null,"external_id":"2611.21423v1","pdf_url":"https://arxiv.org/pdf/2611.21423v1","code_url":"https://github.com/yilish/UniRxnAxis","code_host":"GitHub","authors":["Yili Shen","Kehan Guo","Haomin Zhuang","Xiangliang Zhang"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"Machine learning models for chemical reaction prediction are usually evaluated by endpoint accuracy, but endpoint correctness alone does not reveal how a model represents chemical transformation or why it fails. We introduce UniRxnAxis, a unified forward-retrosynthesis framework that models both tasks as direction-conditioned endpoint redistribution in a shared latent space. By aligning source states, transformed states, and encoded target endpoints, UniRxnAxis enables directional auditing of learned latent updates. This shared latent geometry allows the learned displacement (from the encoder source state $S_0$ to the decoder-facing transformed state $S_1$) to be decomposed into an endpoint-parallel component and an endpoint-orthogonal residual, which we call emergent steering. Our evaluation shows that emergent steering is functionally important: in retrosynthesis, it recovers most of the model's utility, remains effective under random, shuffled, and prototype controls, and directly affects which bond edits are promoted or suppressed. The residual-over-parallel ordering persists across alignment strengths, after decoder-interface calibration, against angle-matched random axes, and in a second graph-edit architecture, although its magnitude depends on the reference axis and optimization. The decomposition also supports a proxy taxonomy of model-side failures, showing how reaction models can be audited beyond endpoint accuracy.","source_metadata":{"categories":["cs.LG","cs.AI"],"code_status":"found","code_url":"https://github.com/yilish/UniRxnAxis"}},{"id":"preprints:2611.21308v1","kind":"preprints","source":"arXiv","title":"Rotation-Invariant Machine Learning for Highly Symmetric Molecular Configurations","url":"https://arxiv.org/abs/2611.21308v1","detail_url":"/chemradar/article?u=https%3A%2F%2Farxiv.org%2Fabs%2F2611.21308v1","date":"2026-11-16T16:06:47Z","timestamp":1794845207,"categories":[],"topic_ids":[],"keywords":["equivariant"],"matched_keywords":["Machine Learning","equivariant","Molecular"],"matched_tags":["machine_learning","equivariance","generic_chem"],"doi":null,"external_id":"2611.21308v1","pdf_url":"https://arxiv.org/pdf/2611.21308v1","code_url":null,"code_host":null,"authors":["Logan Bolton","Tomas Karella","Nicholas Lubbers","Nathaniel Philip Gorski","Emily Shinkle","Alice Elisabeth Anastasia Allen","Pieter J Swart","Roxana Bujack"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"Atomistic machine learning can approximate expensive quantum mechanical calculations at a fraction of their computational cost, but accuracy depends strongly on how the three-dimensional atomic environments are represented. Invariant representations enable fast computation but can fail to distinguish geometrically distinct configurations, especially in highly symmetric environments. We introduce a lightweight benchmark to evaluate expressivity on $n$-fold symmetric configurations, and use it to expose degeneracies in both invariant and equivariant architectures. Guided by these results, we construct a new invariant basis for HIP-HOP which resolves most of the identified failures. On a methane simulation dataset, the extended basis achieves the lowest error among the HIP-HOP variants and has similar accuracy to EquiformerV3 while running $3.2\\times$ faster at inference. These results show that carefully constructed invariant representations can mitigate key expressivity limitations while balancing accuracy and computational efficiency.","source_metadata":{"categories":["physics.chem-ph","cs.CV"],"code_status":"absent"}},{"id":"preprints:2611.21276v1","kind":"preprints","source":"arXiv","title":"Classical structural priors improve neural molecular representations in molecule language models","url":"https://arxiv.org/abs/2611.21276v1","detail_url":"/chemradar/article?u=https%3A%2F%2Farxiv.org%2Fabs%2F2611.21276v1","date":"2026-11-16T15:45:48Z","timestamp":1794843948,"categories":["Cheminformatics","Property Prediction","Design de novo"],"topic_ids":["cheminformatics","property","denovo"],"keywords":["molecule generation","chemoinformatics","molecular representations"],"matched_keywords":["molecule generation","chemoinformatics","molecular representations","molecular","molecule","chemical"],"matched_tags":["de_novo_design","cheminformatics","molecular_descriptor","generic_chem"],"doi":"10.1021/acs.jcim.6c02225","external_id":"2611.21276v1","pdf_url":"https://arxiv.org/pdf/2611.21276v1","code_url":"https://github.com/February24-Lee/SAGA","code_host":"GitHub","authors":["Donghyeok Lee","Sangkwon Park","Munyoung Chang","Sungroh Yoon"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"Classical molecular similarity is central to chemical information systems, yet molecule-language models trained on paired structure-text data do not necessarily organize molecular embeddings according to graded structural relationships. Here we introduce Structure-Aware Guided Alignment (SAGA), a training-only regularizer that uses Centered Kernel Alignment to align neural molecular representations with fingerprint-based Tanimoto similarity, without adding parameters or requiring fingerprints at inference. In MolCA and Atomas, SAGA improves molecule-text retrieval, shifts non-exact retrieval toward closer structural analogues, and increases average scaffold-split property-prediction performance, while molecular description and text-to-molecule generation are improved or preserved in single-run ablations. These results show that established chemoinformatics priors can make molecule-language representations more chemically faithful while retaining their multimodal utility.","source_metadata":{"categories":["cs.LG","cs.CL"],"code_status":"found","code_url":"https://github.com/February24-Lee/SAGA"}},{"id":"preprints:2611.21196v1","kind":"preprints","source":"arXiv","title":"Determining Concentration from Raman Spectroscopy with Manifold Learning and Y-shaped Conformal Autoencoders","url":"https://arxiv.org/abs/2611.21196v1","detail_url":"/chemradar/article?u=https%3A%2F%2Farxiv.org%2Fabs%2F2611.21196v1","date":"2026-11-16T14:52:04Z","timestamp":1794840724,"categories":["Design de novo","Spectra & Analytical"],"topic_ids":["denovo","chemometrics"],"keywords":["gradient boosting","Autoencoder"],"matched_keywords":["process analytical technology","Neural Network","neural networks","gradient boosting","Autoencoder","chemical"],"matched_tags":["process_analytical_technology","neural_network","ensemble","vae","generic_chem"],"doi":null,"external_id":"2611.21196v1","pdf_url":"https://arxiv.org/pdf/2611.21196v1","code_url":null,"code_host":null,"authors":["Anna Taka","Eleni D. Koronaki","Ioannis G. Kevrekidis","Eike Cramer","Joern Viell","Alexander Mitsos"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"Monitoring of chemical processes is crucial in chemical engineering, motivating sophisticated process analytical technology (PAT). In-line analytics are advantageous, with Raman spectroscopy being an established and reliable tool in various applications, from which many properties can be determined. We recently [Koronaki et al., AIChEJ 2024] proposed determining polymer particle size from Raman spectroscopic data, by combining Diffusion Maps (a manifold learning method) with a Y-shaped Conformal Autoencoder Neural Network. Herein, we extend this workflow to predict liquid concentrations, allowing also for multiple target predictions from the same model. We apply the methods on two datasets consisting of Raman measurements and concentration measurements: one from microgel polymerization and one from citric acid dissociation. The single-target concentration prediction results, where one target is determined from each model at a time, are accurate. Comparing against dedicated methods such as Indirect Hard Modeling (IHM) is not straightforward, due to the different validation strategies and the considered spectra range; nevertheless, it is fair to say that IHM performs slightly better than the proposed workflow. We further compare this approach with other regression methods, such as artificial neural networks, extreme gradient boosting and partial least squares, and the results show that the Conformal Autoencoders achieve higher predictive performance, demonstrating that the workflow can also be used for concentration determination. Finally we find that single-target concentration predictions are more accurate than the multiple target predictions for both case studies.","source_metadata":{"categories":["cs.LG"],"code_status":"absent"}},{"id":"preprints:2611.21175v1","kind":"preprints","source":"arXiv","title":"A COMPASS to the Galaxy: Joint nucleosynthetic yield set comparison and parameter inference in Galactic chemical evolution","url":"https://arxiv.org/abs/2611.21175v1","detail_url":"/chemradar/article?u=https%3A%2F%2Farxiv.org%2Fabs%2F2611.21175v1","date":"2026-11-16T14:34:49Z","timestamp":1794839689,"categories":["Design de novo"],"topic_ids":["denovo"],"keywords":["score based","diffusion model","transformer","attention mechanism"],"matched_keywords":["score-based","diffusion model","transformer","attention mechanism","chemical"],"matched_tags":["diffusion","transformer","generic_chem"],"doi":null,"external_id":"2611.21175v1","pdf_url":"https://arxiv.org/pdf/2611.21175v1","code_url":"https://github.com/bGuenes/COMPASS","code_host":"GitHub","authors":["Berkay Günes","Giuseppe Viterbo","Sven Buder","Tobias Buck"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"Galactic Chemical Evolution (GCE) models connect stellar nucleosynthesis to observed abundance patterns, but their interpretation requires simultaneous parameter inference and comparison between competing stellar yield prescriptions. This remains challenging due to intractable likelihoods and growing dataset complexity. Using compass, a novel simulation-based inference framework, we perform, for the first time, amortized joint Bayesian model comparison and parameter inference for GCE models with intractable likelihoods within a unified architecture. Compass employs a conditional score-based diffusion model with a transformer architecture to learn both the posterior and likelihood directly from forward simulations, enabling efficient inference and likelihood-based ranking of competing nucleosynthetic yield sets. In controlled tests, compass identifies the generating yield set and recovers ground-truth parameters with accuracy comparable to existing simulation-based inference (SBI) methods, while reducing computational cost by orders of magnitude. Applied to Milky Way data, the framework favors a hybrid NuGrid + IllustrisTNG yield prescription, confirmed by an independent classifier, and infers an initial mass function slope of $α_{\\rm IMF} = -2.65 \\pm 0.02$ and a SN$_{\\rm Ia}$ normalization of $\\log(N_{\\rm Ia}) = -2.81 \\pm 0.02$, with a systematic spread of $\\approx 0.3$ in $α_{\\rm IMF}$ between the best-ranked yield sets, consistent with independent constraints. The attention mechanism confirms that the inferred parameters are driven by physically meaningful abundance signatures. By unifying amortized parameter inference, Bayesian model comparison, and interpretability in a single probabilistic framework, compass provides a scalable approach for precision chemical evolution studies with current and forthcoming spectroscopic surveys.","source_metadata":{"categories":["astro-ph.GA","astro-ph.IM","cs.LG","physics.data-an"],"code_status":"found","code_url":"https://github.com/bGuenes/COMPASS"}},{"id":"preprints:2611.21116v1","kind":"preprints","source":"arXiv","title":"Out-of-training properties in frozen interatomic potential embeddings are decodable and ordered by physical scale","url":"https://arxiv.org/abs/2611.21116v1","detail_url":"/chemradar/article?u=https%3A%2F%2Farxiv.org%2Fabs%2F2611.21116v1","date":"2026-11-16T13:59:10Z","timestamp":1794837550,"categories":["Property Prediction","ML Potentials"],"topic_ids":["property","potentials"],"keywords":["MLIP","XGBoost","equivariant"],"matched_keywords":["MLIP","XGBoost","equivariant"],"matched_tags":["mlip_tools","ensemble","equivariance"],"doi":null,"external_id":"2611.21116v1","pdf_url":"https://arxiv.org/pdf/2611.21116v1","code_url":"https://github.com/Natalija-Stepurko/mlip-embedding-probes","code_host":"GitHub","authors":["Natalija Stepurko","Joonatan E. M. Laulainen"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"Machine-learned interatomic potentials (MLIPs) are trained on energies, forces and stresses, yet their frozen representations can predict other materials properties. We probe ORB-v3 and UMA-S across layers, read-outs and equivariant channels for nine properties of 154,875 Materials Project crystals. Both models outperform composition-only descriptors on all targets, including five outside their training objectives; errors in physical units help assess potential screening applications. Optimal depth varies by property, consistent with a physical-scale ordering. Random-weight controls reveal strongly property-dependent benefits of pretraining. UMA-S's rotation-invariant angular channels improve crystal-system classification, while higher scores from its full tensor disappear under rotation; matched-model controls do not isolate equivariance as the explanation. Accuracy varies across different probes, but nearest-neighbour prediction often approaches XGBoost performance, even though materials do not form distinct clusters according to their properties. We also test whether a linear predictor trained on one model's embeddings can predict properties using the other model's embeddings after a learned linear transformation. Predictive accuracy is largely preserved, despite differences in which atoms the two models place close together. These results guide the reuse of frozen MLIP representations for materials prediction.","source_metadata":{"categories":["physics.chem-ph","cond-mat.mtrl-sci","cs.LG"],"code_status":"found","code_url":"https://github.com/Natalija-Stepurko/mlip-embedding-probes"}},{"id":"preprints:2611.21079v1","kind":"preprints","source":"arXiv","title":"SemiNovo: Learning Beyond Search-Identified Spectra for De Novo Peptide Sequencing","url":"https://arxiv.org/abs/2611.21079v1","detail_url":"/chemradar/article?u=https%3A%2F%2Farxiv.org%2Fabs%2F2611.21079v1","date":"2026-11-16T13:28:41Z","timestamp":1794835721,"categories":[],"topic_ids":[],"keywords":[],"matched_keywords":["deep learning","chemical","protein"],"matched_tags":["machine_learning","generic_chem","biology"],"doi":null,"external_id":"2611.21079v1","pdf_url":"https://arxiv.org/pdf/2611.21079v1","code_url":"https://github.com/grandOrgan/Seminovo","code_host":"GitHub","authors":["Pan Liu","Dongxin Lyu","Yusen Tan","Yifan Li","Hai-tao Yu","Yunhua Zhong","Zikun Wang","Tianze Ling","Jun Xia"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"Sensitive peptide identification is fundamental to protein characterization and novel protein discovery, for which de novo peptide sequencing (DNPS) from tandem mass spectra provides an important database-independent approach. Although deep learning has substantially advanced de novo sequencing, existing models are predominantly trained on high-confidence peptide-spectrum matches obtained through database search. This search-conditioned supervision excludes a large fraction of unidentified spectra, creating a distribution mismatch between the selectively labeled training data and the complete spectrum stream encountered at inference time, thereby limiting peptide identification performance. To address this challenge, we propose SemiNovo, a semi-supervised framework that learns jointly from identified and unlabeled spectra. SemiNovo first constructs DarkSpec, a large-scale quality-controlled corpus of unlabeled spectra collected from diverse public proteomics projects, thereby extending training beyond search-identified spectra and mitigating the resulting shift in DNPS supervision. It then combines stepwise mass-constrained autoregressive decoding with cumulative path confidence to derive reliable supervision from unlabeled spectra. Moreover, SemiNovo introduces head-wise chemical-evidence gating to improve spectral evidence aggregation under noisy and incomplete fragmentation. Experiments show that SemiNovo performs favorably against state-of-the-art methods on multiple benchmarks. Code and dataset are publicly available.","source_metadata":{"categories":["cs.LG","cs.AI"],"code_status":"found","code_url":"https://github.com/grandOrgan/Seminovo"}},{"id":"feeds:https://www.drugdiscovery.net/2026/11/16/slides-from-ml-in-drug-discovery-symposium-at-the-broad-16-nov/?utm_source=rss&utm_medium=rss&utm_campaign=slides-from-ml-in-drug-discovery-symposium-at-the-broad-16-nov","kind":"feeds","source":"DrugDiscovery.NET","title":"Slides from ML in Drug Discovery Symposium at the Broad, 16 Nov","url":"https://www.drugdiscovery.net/2026/11/16/slides-from-ml-in-drug-discovery-symposium-at-the-broad-16-nov/?utm_source=rss&utm_medium=rss&utm_campaign=slides-from-ml-in-drug-discovery-symposium-at-the-broad-16-nov","detail_url":"/chemradar/article?u=https%3A%2F%2Fwww.drugdiscovery.net%2F2026%2F11%2F16%2Fslides-from-ml-in-drug-discovery-symposium-at-the-broad-16-nov%2F%3Futm_source%3Drss%26utm_medium%3Drss%26utm_campaign%3Dslides-from-ml-in-drug-discovery-symposium-at-the-broad-16-nov","date":"2026-11-16T11:45:13+00:00","timestamp":1794829513,"categories":["Blog"],"topic_ids":[],"keywords":[],"matched_keywords":["ML","Drug"],"matched_tags":["machine_learning","generic_chem"],"doi":null,"external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":[],"journal":null,"publisher":null,"impact_factor":null,"abstract":"","source_metadata":{"source":"DrugDiscovery.NET","published_utc":"2026-11-16T11:45:13+00:00","seen_at":"2026-11-16T14:43:18.911737+00:00"}},{"id":"preprints:2611.20848v1","kind":"preprints","source":"arXiv","title":"MolReps: Taking SELFIES, Making SMILES, Teaching LLMs","url":"https://arxiv.org/abs/2611.20848v1","detail_url":"/chemradar/article?u=https%3A%2F%2Farxiv.org%2Fabs%2F2611.20848v1","date":"2026-11-16T10:23:23Z","timestamp":1794824603,"categories":["Cheminformatics","Property Prediction","LLMs & Agents"],"topic_ids":["cheminformatics","property","llm"],"keywords":["SELFIES","SMILES","InChIKey","LLMs","property prediction","molecular representation","molecular property"],"matched_keywords":["SELFIES","SMILES","InChIKey","LLMs","property prediction","molecular representation","molecule","molecular","molecular property"],"matched_tags":["molecular_notation","LLMs","ml_task","molecular_descriptor","generic_chem","bioactivity"],"doi":null,"external_id":"2611.20848v1","pdf_url":"https://arxiv.org/pdf/2611.20848v1","code_url":null,"code_host":null,"authors":["Ken Deng","Garrett Morris"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"The same molecule can be expressed in multiple, syntactically distinct representations, yet it remains unclear whether large language models (LLMs) can recognize and relate these equivalent encodings. We introduce MolReps, a pipeline that generates 9,600 question-answer pairs across ten representations for conversion, normalization, and 2D structure recognition. Post-training on MolReps improves text-only and multimodal models, MolQwen3-4B-Instruct (MQ3-4B-I) and MolQwen3-VL-4B-Instruct (MQ3-VL-4B-I), from $4.01\\%$ to $75.13\\%$ and from $7.11\\%$ to $73.14\\%$ (InChIKey match) on the molecule-overlap test set, respectively. On the molecule-disjoint test set, post-training increased validity from $20.35\\%$ to $58.12\\%$ for MQ3-4B-I and from $17.19\\%$ to $59.34\\%$ for MQ3-VL-4B-I. However, InChIKey match remained low, decreasing from $3.86\\%$ to $3.09\\%$ and from $6.44\\%$ to $2.29\\%$, respectively. Furthermore, the resulting models achieve a relative $27.84\\%$ improvement in average over their base model on MolRepBench, a comprehensive benchmark for evaluating molecular property prediction. These results show the effectiveness of targeted post-training for enhancing molecular representation understanding and its potential to improve downstream molecular tasks.","source_metadata":{"categories":["physics.chem-ph","cs.AI","cs.CL"],"code_status":"absent"}},{"id":"preprints:2611.20706v1","kind":"preprints","source":"arXiv","title":"Adsorbate-Free Screening of Dissociative Water Adsorption on Nickel Nanoparticles via Graph Neural Networks","url":"https://arxiv.org/abs/2611.20706v1","detail_url":"/chemradar/article?u=https%3A%2F%2Farxiv.org%2Fabs%2F2611.20706v1","date":"2026-11-16T08:13:59Z","timestamp":1794816839,"categories":["Property Prediction","ML Potentials"],"topic_ids":["property","potentials"],"keywords":["MACE","Graph Neural Networks","graph neural network","GNN","GNNs","density functional theory","DFT"],"matched_keywords":["MACE","Neural Networks","neural network","Graph Neural Networks","graph neural network","GNN","GNNs","density functional theory","DFT"],"matched_tags":["mlip_tools","neural_network","gnn","dft"],"doi":null,"external_id":"2611.20706v1","pdf_url":"https://arxiv.org/pdf/2611.20706v1","code_url":"https://github.com/mmonti00/adsorbate-free_GNN_Ni_NPs","code_host":"GitHub","authors":["Marta Monti","Alec Noppe","Olivier Lugier","Maria B. Brands","Bernd Ensing"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"Dissociative water adsorption is a key step in heterogeneous and electrochemical catalysis, but on finite Ni nanosystems the OH$^*$ and H$^*$ fragments can occupy many combinations of surface sites, making exhaustive density functional theory (DFT) screening impractical. We introduce a graph neural network (GNN) framework with an adsorbate-free representation: candidate O- and H-binding sites are defined by sets of Ni atoms (\\textit{anchors}), and each site pair is encoded as a graph constructed solely from these anchors and their local Ni environment. Using only MACE-relaxed bare Ni geometries as input, separate GNNs predict the DFT-reference adsorption energies of OH$^*$+H$^*$ configurations and whether the proposed binding sites would be retained upon DFT relaxation, eliminating the need to construct or relax adsorbates during screening. Trained across Ni nanoclusters and Ni(111)/Ni(100) surfaces, the energy model achieves a mean absolute error of 0.031 eV on the internal validation set and demonstrates transferability to unseen Ni$_{55}$, Ni$_{91}$, and Ni$_{130}$ clusters and larger surface cells, particularly for relative-energy ranking, although quantitative accuracy varies across systems. These independent tests also support the retention classifier's use as a conservative filter for selecting configurations likely to preserve their proposed binding sites. Application to Ni nanoparticles containing up to $\\sim$3000 atoms identifies favorable co-adsorption motifs and shows that lower O-site coordination generally stabilizes the dissociated state, whereas H-site coordination has a weaker influence. The framework thus offers an efficient route to screening adsorption energetics, with potential extensions to other metals and adsorbate combinations.","source_metadata":{"categories":["cond-mat.mtrl-sci","physics.chem-ph"],"code_status":"found","code_url":"https://github.com/mmonti00/adsorbate-free_GNN_Ni_NPs"}},{"id":"preprints:2611.20603v1","kind":"preprints","source":"arXiv","title":"BiGSyn: A Bi-Level Heterogeneous Graph Contrastive Learning Framework for Interpretable Drug Synergy Prediction","url":"https://arxiv.org/abs/2611.20603v1","detail_url":"/chemradar/article?u=https%3A%2F%2Farxiv.org%2Fabs%2F2611.20603v1","date":"2026-11-16T06:28:13Z","timestamp":1794810493,"categories":["ADMET & Safety"],"topic_ids":["admet"],"keywords":["transformers","attention mechanism"],"matched_keywords":["transformers","attention mechanism","Drug","molecular"],"matched_tags":["transformer","generic_chem"],"doi":null,"external_id":"2611.20603v1","pdf_url":"https://arxiv.org/pdf/2611.20603v1","code_url":"https://github.com/senquanwang/BiGSyn","code_host":"GitHub","authors":["Senquan Wang","Hui Fang","Haoran Nan","Xingwei Wang","Qiang He"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"Drug combination therapy has become a mainstream of modern medicine, with the potential to enhance efficacy, reduce toxicity, and overcome drug resistance. However, identifying synergistic drug combinations remains challenging, as it requires modeling complex interactions across multiple biological scales---from atom-level molecular structures to entity-level relations in biological networks. Existing approaches typically model these levels in isolation, overlooking complementary signals and long-range dependencies. To address these limitations, we propose BiGSyn, a Bi-level heterogeneous Graph contrastive learning framework for drug Synergy prediction. Specifically, BiGSyn leverages heterogeneous graph transformers to extract high-order dependencies from molecular graphs and biological networks, and introduces a bi-level contrastive learning objective to bridge the cross-level semantic gap. Extensive experiments on two benchmark datasets show that BiGSyn consistently outperforms state-of-the-art baselines. Moreover, its attention mechanism highlights key molecular substructures and biological interactions driving synergy, providing interpretable insights.","source_metadata":{"categories":["cs.AI"],"code_status":"found","code_url":"https://github.com/senquanwang/BiGSyn"}},{"id":"preprints:2611.20429v1","kind":"preprints","source":"arXiv","title":"FiuMD: Heterogeneous Batching for Coarse-Grained Molecular Dynamics with Neural Network Potentials","url":"https://arxiv.org/abs/2611.20429v1","detail_url":"/chemradar/article?u=https%3A%2F%2Farxiv.org%2Fabs%2F2611.20429v1","date":"2026-11-16T02:44:32Z","timestamp":1794797072,"categories":["ML Potentials"],"topic_ids":["potentials"],"keywords":["Neural Network Potentials","neural network potential","Molecular Dynamics"],"matched_keywords":["Neural Network Potentials","neural network potential","force-field parameters, periodic box, timestep, or thermostat parameters","Neural Network","Molecular","Molecular Dynamics","protein"],"matched_tags":["mlip_tools","force_field_parameterization","neural_network","generic_chem","molecular_dynamics","biology"],"doi":null,"external_id":"2611.20429v1","pdf_url":"https://arxiv.org/pdf/2611.20429v1","code_url":"https://github.com/hernandezhdd/fiumd","code_host":"GitHub","authors":["Hugo D. Hernández R.","Paola B. Torres","Lara Negri","Rocío L. Cecchini","Nicolás A. García"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"FiuMD is a GPU molecular dynamics engine that advances many independent biomolecular simulations as a single batch, even when the systems differ in size, topology, force-field parameters, periodic box, timestep, or thermostat parameters. All systems are concatenated into global tensors, so the classical interaction terms and a shared TorchMD-Net neural network potential are evaluated for the whole batch at once, while each system keeps its own integration and thermostat. FiuMD started as a fork of TorchMD and reworks its system representation, force evaluation, and integration to handle such heterogeneous batches. We test it with coarse-grained C$α$ models of chignolin, Trp-cage, BBL, and $λ$-repressor. In double precision, forces within a heterogeneous batch agree with isolated FiuMD and TorchMD evaluations to within the precision of repeated evaluations, with a maximum relative difference of $7\\times10^{-16}$ across 60 systems. Over 1000 steps without a thermostat, batched trajectories deviate by at most $2.5\\times10^{-11}$ Angstrom from isolated ones, even when periodic images lie within the neural network cutoff. With a different target temperature for each system of the ensemble, the mean and width of each system's kinetic temperature match the values expected from its own thermostat. We show that when individual simulations are too small to fill the GPU, batching can improve throughput several-fold over TorchMD. On a single GPU, batching four different protein systems achieves a 4.6-7.8-fold speedup per step compared with sequential TorchMD simulations. Compared with sequential homogeneous FiuMD batches, heterogeneous batching provides up to a 4-fold speedup for smaller batches while maintaining comparable performance for larger ones.","source_metadata":{"categories":["physics.comp-ph","cond-mat.soft"],"code_status":"found","code_url":"https://github.com/hernandezhdd/fiumd"}},{"id":"journals:3169082f87992e6c59e618e36b21b0ea1995bce7","kind":"journals","source":"Molecules","title":"A Bounded Design Space for Ionic Liquid-Assisted Extraction of Essential Oils from Seven Medicinal Plants: Chemical Profiling and Antibacterial Potential of Amomum villosum Lour. Essential Oil","url":"https://doi.org/10.3390/molecules31224134","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.3390%2Fmolecules31224134","date":"2026-11-16T00:00:00Z","timestamp":1794787200,"categories":[],"topic_ids":[],"keywords":["regression model","bioactivity"],"matched_keywords":["regression model","Chemical","bioactivity"],"matched_tags":["ml_task","generic_chem","bioactivity"],"doi":"10.3390/molecules31224134","external_id":"3169082f87992e6c59e618e36b21b0ea1995bce7","pdf_url":null,"code_url":null,"code_host":null,"authors":["G. Shi","Yun Xiang","Ya-Juan Lin","Hong-Juan Bao"],"journal":"Molecules","publisher":null,"impact_factor":null,"abstract":"Essential oils are promising sources of antimicrobial and pharmaceutical agents, but process optimization still follows a “one medicinal plant–one process model” paradigm. Drawing on the core philosophy of Quality by Design, this work developed a raw-material-aware bounded design space for microwave-assisted ionic-liquid pretreatment combined with hydro-distillation (MILT-HD) across seven medicinal plants. A 22-run D-optimal design integrated three process variables with plant microstructural descriptors characterized by mercury intrusion porosimetry and cell-wall measurement. Multi-block partial least squares regression linked porous matrix features, cell wall thickness and operational parameters to an extraction performance index (EPI). Microwave time and power were key process parameters, and cell wall thickness was a key material attribute. The three-variable multiple linear regression model explained 72% of EPI variance and defined a design space targeting EPI > 3. In-domain validation and external validation on unseen species gave absolute relative errors of 2.82–6.41% and 6.75–8.57%, respectively. Monte Carlo simulation confirmed a robust operating region with failure probability <1%. Using Amomum villosum Lour. as the case study, MILT-HD raised oil yield from 28.9 to 32.8 µL·g−1 and shortened extraction from 3.5 h to 1.1 h while largely preserving the characteristic volatile profile. The obtained oil inhibited Salmonella enteritidis, disrupting bacterial envelope integrity and reducing intracellular ATP. This cross-species bounded design space is beneficial for reducing repeated optimization workload for materials with comparable microstructural attributes and providing a data-driven strategy to produce high-quality essential oils for pharmaceutical bioactivity exploration.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:232c46f6d080ce1e7ebe643f3f912ea780cbc36b","kind":"journals","source":"IOASD Journal of Medical and Pharmaceutical Sciences","title":"A Comprehensive Review on Molecular Mechanism of HIF-1 alpha Pathway in Breast Cancer","url":"https://doi.org/10.67224/ioasdjmps.2026.v03i04.014","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.67224%2Fioasdjmps.2026.v03i04.014","date":"2026-11-16T00:00:00Z","timestamp":1794787200,"categories":["Docking & Screening"],"topic_ids":["docking"],"keywords":["Molecular docking","receptor"],"matched_keywords":["Molecular docking","Molecular","drug","receptor"],"matched_tags":["docking","generic_chem","biology"],"doi":"10.67224/ioasdjmps.2026.v03i04.014","external_id":"232c46f6d080ce1e7ebe643f3f912ea780cbc36b","pdf_url":null,"code_url":null,"code_host":null,"authors":["D. Manimaran","K. Ramakrishnan","Kowsalya Rengasamy","N. Elangovan"],"journal":"IOASD Journal of Medical and Pharmaceutical Sciences","publisher":null,"impact_factor":null,"abstract":"Breast cancer is one of the most prevalent malignancies worldwide, characterized by uncontrolled cell proliferation, angiogenesis, invasion, and metastasis. Tumor hypoxia contributes significantly to breast cancer progression through the activation of hypoxia-inducible factor-1 alpha (HIF-1α), which regulates cellular adaptation to low oxygen conditions. This review highlights the molecular mechanisms of HIF-1α signaling and its involvement in tumor progression, angiogenesis, metastasis, and therapeutic resistance. Conventional treatment strategies, including chemotherapy, radiotherapy, and hormone therapy, are discussed alongside emerging peptide-based therapeutic approaches. Particular attention is given to peptide-based hormone therapy, receptor-targeted peptides, peptide receptor radionuclide therapy, peptide vaccines, anticancer peptides, and peptide-mediated drug delivery. Key molecular targets, including epidermal growth factor receptor (EGFR), vascular endothelial growth factor (VEGF), and HIF-1α, are examined. The review further explores hypoxia-associated metabolic adaptations, transcriptional regulation, apoptosis, cell proliferation, and tumor survival pathways. Molecular docking is highlighted as a promising approach for identifying potential anticancer therapeutic candidates. Additionally, advanced peptide delivery systems involving cell-penetrating peptides, chitosan, sodium alginate, biocompatible polymers, and nanoparticles are discussed for improving targeted delivery and therapeutic efficacy. Animal models are also considered for evaluating treatment strategies. Overall, integrating HIF-1α-targeted interventions with peptide-based drug delivery may offer promising opportunities for developing innovative breast cancer therapies.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:d79cf9e549b5398fe2ef5a57f04f91f7acf5017a","kind":"journals","source":"Nature protocols","title":"A robotics- and machine learning-integrated workflow for discovering all-natural plastic substitutes","url":"https://doi.org/10.1038/s41596-026-01449-4","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1038%2Fs41596-026-01449-4","date":"2026-11-16T00:00:00Z","timestamp":1794787200,"categories":[],"topic_ids":[],"keywords":["molecular dynamics"],"matched_keywords":["neural network","machine learning","molecular","molecular dynamics"],"matched_tags":["neural_network","machine_learning","generic_chem","molecular_dynamics"],"doi":"10.1038/s41596-026-01449-4","external_id":"d79cf9e549b5398fe2ef5a57f04f91f7acf5017a","pdf_url":null,"code_url":null,"code_host":null,"authors":["Yang Li","Abhishek T. Sose","Tian-Le Chen","Po-Yen Chen"],"journal":"Nature protocols","publisher":null,"impact_factor":null,"abstract":"The global accumulation of nondegradable petrochemical plastic waste is driving the search for biodegradable substitutes derived from natural resources. However, discovering all-natural composite materials that match the optical, thermal and mechanical performance of commercial plastics remains challenging because traditional iterative optimization is inefficient. Here the workflow described in this Protocol accelerates the screening process by using automated robotics for liquid handing and machine learning to predict material properties. The procedure begins by commanding a pipetting robot to prepare a library of 286 nanocomposites. Basic physical properties of these products are used to train a support vector machine classifier that locates the decision boundary separating feasible from infeasible film compositions in the four-component system, thereby defining the feasible design space. Next, through 14 active learning loops coupled with data augmentation (for example, adding small random changes to the characterization data within the observed experimental error range), the robot fabricates 135 all-natural nanocomposites in a stagewise manner to construct an artificial neural network prediction model. The model is used to predict (1) the physicochemical properties of a nanocomposite from its composition and (2) which compositions are likely to have the desired, user-specified properties (for example, transparency or fire resistance). This protocol further details methods for model interpretation using Shapley Additive Explanations, validation via molecular dynamics simulations and model expansion to incorporate new building blocks. This hybrid approach takes ~2 months and enables the accelerated discovery of eco-friendly materials with programmable functions using components drawn from the generally recognized as safe database.","source_metadata":{"source":"semantic_scholar"}},{"id":"preprints:10.26434/chemrxiv.15012268/v1","kind":"preprints","source":"ChemRxiv","title":"A vacated second ligand site acts as a decoy in redocking: systematic in P-glycoprotein, intermittent across eight other proteins","url":"https://doi.org/10.26434/chemrxiv.15012268/v1","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.26434%2Fchemrxiv.15012268%2Fv1","date":"2026-11-16T00:00:00Z","timestamp":1794787200,"categories":["Docking & Screening"],"topic_ids":["docking"],"keywords":["rDock","smina","receptor"],"matched_keywords":["rDock","smina","ligand","receptor"],"matched_tags":["docking_software","generic_chem","biology"],"doi":"10.26434/chemrxiv.15012268/v1","external_id":"10.26434/chemrxiv.15012268/v1","pdf_url":null,"code_url":null,"code_host":null,"authors":["Hayat Ullah","Nada Mohamed Yassein Farih","Louisa Ayamga","Zeba khan","Khwaja Jawad","Muhammad Yaseen","Ziaullah Khan","Md Ariful Islam Mozumder","Moon-Il Joo","Hee-Cheol Kim"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"Redocking tests whether a docking protocol can recover a deposited pose: the ligand is removed and the program is asked to put it back. Standard preparation deletes every copy. When a structure contains two copies in one cavity, the receptor passed to the program was never observed empty. We measured the effect of that deletion in deposited human P-glycoprotein structures, with the search box centred on the deposited pose so that a failure cannot be blamed on site finding. Across 15 structures the median top-pose RMSD is 8.30 Å and 1 of 15 falls within 2.0 Å. Accuracy does not track the density support of the deposited ligand. That was the registered hypothesis, and it is not supported (Spearman ρ = −0.004, p = 0.9899). In the nine structures that carry a second copy, retaining the copy as rigid receptor moves the median from 8.26 to 3.46 Å and improves 8 of 9 (p = 0.0078). A carbon-only ghost of the same shape, which the scoring function sees as apolar volume but not as a hydrogen-bonding partner, agrees with the retained copy (p = 0.1289), while a box reduced to the same volume does not (p = 0.2969). The displaced poses lie in the vacated subsite (mean overlap 0.216 to 0.539, p = 0.0078). The comparison reproduces in rDock, which shares no code with smina: 6 of 7 structures improve, and the median falls from 8.35 to 3.67 Å (p = 0.0469). On eight unrelated proteins the same protocol does not support one direction (5 of 8, p = 0.3750), although trypsin moves from 8.59 to 0.50 Å. A scan of 37,596 PDB entries found 4,774 multi-copy sites across 929 proteins. A redocking number for a multi-copy structure should be reported both with the other copies removed and with them retained. Scientific Contribution. Published evidence for this preparation effect was one ligand in one structure. The comparison here covers nine two-copy P-glycoprotein structures, and a carbon ghost together with a volume-matched box separates apolar occlusion from polar ligand chemistry and from the volume searched. Repeated on eight other proteins, the same comparison does not support a consistent direction.","source_metadata":{"date_field":"posted","keywords":[]}},{"id":"journals:4fa47930b1beef97defa1bee94b1907ee809e209","kind":"journals","source":"Frontiers in Biophysics","title":"Advances in prediction of drug-target thermodynamics and kinetics: from molecular simulations to artificial intelligence","url":"https://doi.org/10.3389/frbis.2026.1953438","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.3389%2Ffrbis.2026.1953438","date":"2026-11-16T00:00:00Z","timestamp":1794787200,"categories":[],"topic_ids":[],"keywords":["Molecular Dynamics","MD simulations"],"matched_keywords":["ML","drug","molecular","compounds","compound","drugs","ligand","Molecular Dynamics","MD simulations","protein"],"matched_tags":["machine_learning","generic_chem","molecular_dynamics","biology"],"doi":"10.3389/frbis.2026.1953438","external_id":"4fa47930b1beef97defa1bee94b1907ee809e209","pdf_url":null,"code_url":null,"code_host":null,"authors":["C. Etchebest","A. de Brevern","J. Diharce"],"journal":"Frontiers in Biophysics","publisher":null,"impact_factor":null,"abstract":"In the drug discovery field, drug affinity has long been considered the gold standard for designing small compounds with therapeutic activity. However, since the beginning of the 21st century, kinetic parameters, which are directly related to compound efficacy, have increasingly been recognized as central to the rational design of new drugs. However, such data are scarce, while the experimental determination of the association and dissociation rate constants, k on and k off , is costly and time-consuming. Molecular simulations, such as Molecular Dynamics (MD) calculations have been proposed to predict those kinetic constants, despite known limitations. With the rapid development of AI-based tools, researchers are now proposing advanced methodologies that leverage both data-driven learning from existing datasets and physics-based modelling through MD simulations. This review presents a critical overview of MD-based approaches for estimating protein-ligand binding free energies and rate constants. It then examines data-driven prediction methods and ML-assisted molecular simulations, with particular attention to validation and current limitations. Finally, it discusses the crucial need for new data with respect to the FAIR principles and methodological developments needed to further improve prediction quality.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:e2d9d7067dc6bcaa0b5623794af851c4effb6385","kind":"journals","source":"Phytotherapy research","title":"Alginic Acid Ameliorates Osteoarthritis Through FOXO3a-Dependent Regulation of the EREG-MAPK Signaling Axis","url":"https://doi.org/10.1002/ptr.70501","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1002%2Fptr.70501","date":"2026-11-16T00:00:00Z","timestamp":1794787200,"categories":[],"topic_ids":[],"keywords":["virtual screening"],"matched_keywords":["virtual screening"],"matched_tags":["virtual_screening"],"doi":"10.1002/ptr.70501","external_id":"e2d9d7067dc6bcaa0b5623794af851c4effb6385","pdf_url":null,"code_url":null,"code_host":null,"authors":["R. Liu","Long Chen","Shi-Han Wang","Bing-Qi Pan","Tian-Yu Zhang","Hui Chen","Xiao-Lei Li"],"journal":"Phytotherapy research : PTR","publisher":null,"impact_factor":null,"abstract":"Osteoarthritis (OA) lacks effective disease-modifying therapies. This study evaluated the therapeutic potential of alginic acid (AA), which was identified through a FOXO3a-oriented virtual screening strategy, and investigated its downstream regulatory mechanism with a focus on the EREG-MAPK signaling axis. AA was prioritized through structure-based virtual screening against FOXO3a followed by functional secondary screening. In vitro, IL-1β-stimulated human OA chondrocytes were treated with AA and assessed for extracellular matrix metabolism, inflammation, apoptosis, and MAPK pathway activity. FOXO3a knockdown and EREG-overexpression rescue experiments were performed to evaluate the functional roles of FOXO3a and EREG. In vivo, DMM-induced OA rats received intra-articular AA treatment, followed by assessment of pain behavior, cartilage pathology, and matrix preservation. Transcriptomic analysis was used to identify AA-responsive genes and signaling pathways under OA-like inflammatory conditions. AA promoted cartilage matrix synthesis, reduced matrix degradation and inflammatory responses, and attenuated chondrocyte apoptosis in vitro. In DMM-induced OA rats, AA alleviated pain hypersensitivity and reduced cartilage degeneration. Transcriptomic and subsequent validation analyses identified EREG and MAPK signaling as AA-responsive components. AA reduced EREG expression and MAPK phosphorylation. Importantly, FOXO3a knockdown markedly attenuated the effects of AA on EREG expression, p38 MAPK activation, and cartilage matrix homeostasis, supporting a functional requirement for FOXO3a in AA-mediated chondroprotection. Conversely, EREG overexpression partially counteracted the protective effects of AA. AA exerts chondroprotective effects in experimental OA and appears to act, at least in part, through FOXO3a-dependent suppression of EREG-MAPK signaling. These findings strengthen the mechanistic basis for AA as a potential OA therapeutic candidate while supporting further investigation of the FOXO3a-EREG-MAPK regulatory axis.","source_metadata":{"source":"semantic_scholar"}},{"id":"preprints:10.26434/chemrxiv.15012010/v2","kind":"preprints","source":"ChemRxiv","title":"AnCore: A Six-Stage Structure-Based Virtual Screening Pipeline Integrating Two-Phase Redocking, Reciprocal Rank Fusion Consensus Scoring, and a Contact-Anchored Elastic-Network Pose-Persistence Heuristic","url":"https://doi.org/10.26434/chemrxiv.15012010/v2","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.26434%2Fchemrxiv.15012010%2Fv2","date":"2026-11-16T00:00:00Z","timestamp":1794787200,"categories":["Cheminformatics"],"topic_ids":["cheminformatics"],"keywords":["RDKit","Virtual Screening","AutoDock Vina","receptor"],"matched_keywords":["RDKit","Virtual Screening","AutoDock Vina","molecular","ligand","ligands","receptor"],"matched_tags":["cheminformatics_software","virtual_screening","docking_software","generic_chem","biology"],"doi":"10.26434/chemrxiv.15012010/v2","external_id":"10.26434/chemrxiv.15012010/v2","pdf_url":null,"code_url":null,"code_host":null,"authors":["Norseen H Abdelhakim","Ahmed M. Senary"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"Structure-based virtual screening workflows often keep molecular visualization apart from multi-engine consensus rescoring and post-docking triage. We present AnCore Scientific Studio (v1.2.0), an open-source macOS workbench that couples a Rust/Tauri shell and a 3Dmol.js viewport to a six-stage Python pipeline. The stages are (1) receptor curation with AMBER14/GBn2 minimization and a steric clash gate; (2) a two-phase co-crystal redocking gate, with local BFGS refinement (Phase A) and focused-box Monte Carlo search (Phase B); (3) RDKit ligand preparation; (4) consensus rescoring of AutoDock Vina, Vinardo and AutoDock4 by Reciprocal Rank Fusion and mean Z-score; (5) 3D radial interaction fingerprinting; and (6) post-docking triage, made up of unrestrained full-atom minimization (Mode A), a 3.0 ps contact-anchored elastic-network pose-persistence heuristic (Mode B) and export of 100-ns explicit-solvent OpenMM/GROMACS packages (Mode C). We tested the pipeline on a five-target development benchmark (1IEP, 3ERT, 1UYG, 1M17, 1STP; 50 ligands, 25 literature actives and 25 presumed-inactive decoys, plus 5 co-crystal references). All five co-crystal ligands passed the 2.00 Å redocking gate (mean RMSD 0.906 ± 0.388 Å). Consensus rescoring raised the suite-mean ROC-AUC from 0.584 for Vina alone to 0.696 for mean Z-score (one-sided paired bootstrap p = 0.0070). A contact-count control that uses no co-crystal information reached 0.864, and the Mode B heuristic reached 0.920. Because the Mode B parameters were set during development on these same targets, and because its inputs overlap with the contact counts it is compared against, we treat these two figures as in-sample results. Mode B also rejected 40% of the actives under hard cutoffs, so it should be used for ranking and not for rejection. Prospective validation on independent DUD-E and LIT-PCBA targets is still needed.","source_metadata":{"date_field":"posted","keywords":[]}},{"id":"journals:09332d8bc94194a87ff43826a74558d8b802a993","kind":"journals","source":"The Journal of Physical\nChemistry B","title":"Cationic Gemini\nSurfactants with Cleavable Spacer\nand Tails as Multifunctional Bioactive Platforms: Self-Assembly, DNA\nBinding, Gene Transfection, and Antibacterial Activity Studies","url":"https://doi.org/10.1021/acs.jpcb.6c06529","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1021%2Facs.jpcb.6c06529","date":"2026-11-16T00:00:00Z","timestamp":1794787200,"categories":["Docking & Screening"],"topic_ids":["docking"],"keywords":["molecular docking"],"matched_keywords":["molecular docking","molecular","chemical"],"matched_tags":["docking","generic_chem"],"doi":"10.1021/acs.jpcb.6c06529","external_id":"09332d8bc94194a87ff43826a74558d8b802a993","pdf_url":null,"code_url":null,"code_host":null,"authors":["Homen Dahal","Sadhana Roy","Shashi Kumar","J. Dey","Somdeb BoseDasgupta"],"journal":"The Journal of Physical\nChemistry B","publisher":null,"impact_factor":null,"abstract":"Cleavable surfactants have gained significant interest in surfactant science due to their controlled degradation, improved biocompatibility, and potential applications in biomedical and industrial fields. The development of cleavable Gemini surfactants, which offer superior surface activity and self-assembly behavior compared with conventional single-chain surfactants, provides a promising approach toward multifunctional and biodegradable surfactant systems. Herein, we report the design and synthesis of two novel cationic Gemini surfactants, C14E-EE-EC14 and C14E-AA-EC14, featuring strategically engineered cleavable spacer architectures containing diester and diamide functionalities, respectively. In addition to the cleavable bonds incorporated within the spacer, both surfactants contain ester linkages connecting the hydrophobic tails, resulting in four hydrolytically cleavable functionalities within the molecular framework, which enhance biodegradability and reduce potential toxicity. Their physicochemical properties were systematically investigated using surface tension, conductivity, isothermal titration calorimetry, and fluorescence probe studies. Both surfactants exhibited excellent surface activity and spontaneous self-assembly, while the chemical nature of the spacer linkage strongly influenced their aggregation thermodynamics and microenvironmental properties. Transmission electron microscopic analysis revealed distinct morphology governed by spacer rigidity, where the diester-linked surfactant formed smaller micellar aggregates along with vesicular structures, whereas the corresponding diamide-linked derivative generated elongated aggregates and smaller vesicular assemblies. The cationic Geminis showed efficient DNA interaction, confirmed by ethidium bromide exclusion assays and molecular docking studies. This improved complexation ability resulted in effective gene delivery, with both surfactants demonstrating comparable transfection efficiency to TurboFect in RAW 264.7 cells. Additionally, their surface-active characteristics translated into remarkable antibacterial activity against both Gram-negative and Gram-positive bacteria, comparable to kanamycin and ampicillin. Overall, these cleavable Gemini surfactants represent promising multifunctional platforms combining biodegradability, tunable self-assembly, gene delivery capability, and antimicrobial activity.","source_metadata":{"source":"semantic_scholar"}},{"id":"preprints:10.26434/chemrxiv.15012303/v1","kind":"preprints","source":"ChemRxiv","title":"Chemistry Lab Notebook for Obsidian: An ELN for CDX and CDXML Reaction Files","url":"https://doi.org/10.26434/chemrxiv.15012303/v1","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.26434%2Fchemrxiv.15012303%2Fv1","date":"2026-11-16T00:00:00Z","timestamp":1794787200,"categories":["Cheminformatics"],"topic_ids":["cheminformatics"],"keywords":["SMILES","RDKit","PubChem"],"matched_keywords":["SMILES","RDKit","PubChem"],"matched_tags":["molecular_notation","cheminformatics_software","chem_benchmarks"],"doi":"10.26434/chemrxiv.15012303/v1","external_id":"10.26434/chemrxiv.15012303/v1","pdf_url":null,"code_url":null,"code_host":null,"authors":["Pavel Sakharov","Mark Gandelman"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"We describe the Chemistry Lab Notebook for Obsidian, an electronic laboratory notebook (ELN) for synthetic chemistry. Each Markdown reaction note is linked to a CDX or CDXML file drawn in ChemDraw or in the free Reactio Draw. The notebook's plugins display the reaction scheme and reagent table of this file in the note and search all reactions by structure, role or whole reaction. They also draft characterization paragraphs from the note fields and export Open Reaction Database records and PubChem deposition files. In a controlled test, 10,000 patent reactions were written programmatically as CDXML files and read back without change in connectivity or defined stereochemistry. On these files, all 306 search queries returned the same reactions as RDKit matching on the source SMILES. Seven chemists in our group have recorded about 4000 reaction notes in the notebook.","source_metadata":{"date_field":"posted","keywords":[]}},{"id":"journals:9e77bf811680a363809eff124208f79cbd620d07","kind":"journals","source":"International Journal of Molecular Sciences","title":"Chrysin Inhibits Telomerase and Induces Senescence in Two Cancer Cell Lines: Insights from In Vitro Analysis, Molecular Docking, DFT, and ADMET","url":"https://doi.org/10.3390/ijms272210282","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.3390%2Fijms272210282","date":"2026-11-16T00:00:00Z","timestamp":1794787200,"categories":["Docking & Screening","ADMET & Safety"],"topic_ids":["docking","admet"],"keywords":["Molecular Docking","DFT","density functional theory","ADMET"],"matched_keywords":["Molecular Docking","DFT","density functional theory","ADMET","Molecular","protein"],"matched_tags":["docking","dft","admet","generic_chem","biology"],"doi":"10.3390/ijms272210282","external_id":"9e77bf811680a363809eff124208f79cbd620d07","pdf_url":null,"code_url":null,"code_host":null,"authors":["Mahendar Porika","Radhika Tippani","Sirisha Kalam","D. Banothu","Sriram Rekulapally","R. R. Poonuru","S. Abbagani","G. Saretzki"],"journal":"International Journal of Molecular Sciences","publisher":null,"impact_factor":null,"abstract":"Telomerase activity (TA) is crucial for preserving telomeres and preventing senescence, while telomerase inhibition can cause cellular senescence in cancer cells for therapeutic applications. Certain phytochemical substances act as telomerase inhibitors, but understanding their impact on lung and cervical cancer cells is limited. Our study investigated the effects of the natural polyphenol chrysin on A549 and HeLa cell senescence, as well as its interaction with telomerase in vitro and in silico. Chrysin displayed no red blood cell cytotoxicity on or viability changes in WI-38 fibroblasts up to 60 μM. Chrysin decreased TA and cell proliferation at early time points and induced senescence dose-dependently at a later time point in both cell lines. Molecular docking studies demonstrated robust interaction with the telomerase protein TERT, resulting in a Piecewise Linear Potential (PLP) score of −53.79. A density functional theory (DFT) experiment revealed critical insights into molecular characteristics like net charge, orbital geometries (HOMO and LUMO), energy gap (ΔE), and dipole moment. Quantum parameters and ADMET (absorption, distribution, metabolism, elimination, and toxicity) properties of chrysin suggest pharmacological properties and safety. Our findings underscore the intricate interplay between telomerase inhibition, anti-proliferative activity, and senescence, offering valuable insights for future development of novel telomerase-based therapies.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:10.1021/acs.jcim.6c02225","kind":"journals","source":"Journal of Chemical Information and Modeling","title":"Classical Structural\nPriors Improve Neural Molecular\nRepresentations in Molecule Language Models","url":"https://doi.org/10.1021/acs.jcim.6c02225","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1021%2Facs.jcim.6c02225","date":"2026-11-16T00:00:00+00:00","timestamp":1794787200,"categories":["Cheminformatics","Property Prediction","Design de novo"],"topic_ids":["cheminformatics","property","denovo"],"keywords":["molecule generation","chemoinformatics","property prediction","molecular representations"],"matched_keywords":["molecule generation","chemoinformatics","property prediction","molecular representations","Molecular","Molecule","chemical"],"matched_tags":["de_novo_design","cheminformatics","ml_task","molecular_descriptor","generic_chem"],"doi":"10.1021/acs.jcim.6c02225","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Donghyeok Lee","Sangkwon Park","Munyoung Chang","Sungroh Yoon"],"journal":"Journal of Chemical Information and Modeling","publisher":"ACS","impact_factor":5.3,"abstract":"Classical molecular similarity is central to chemical information systems, yet molecule-language models trained on paired structure–text data do not necessarily organize molecular embeddings according to graded structural relationships. Here, we introduce Structure-Aware Guided Alignment (SAGA), a training-only regularizer that uses Centered Kernel Alignment to align neural molecular representations with fingerprint-based Tanimoto similarity and requires no additional parameters or fingerprints at inference. In MolCA and Atomas, SAGA improves molecule–text retrieval, retrieves closer structural analogues, and raises average performance in scaffold-split property prediction, without clear degradation of molecular description or text-to-molecule generation. These results, together with control experiments that attribute the gains to the chemical content of the prior, show that established chemoinformatics priors can make molecule-language representations more chemically faithful while retaining their multimodal utility.","source_metadata":{"short_name":"JCIM","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"journals:f26770a898f942ae759c60a12c91319e2fffb042","kind":"journals","source":"Journal of Chemical Information\nand Modeling","title":"Classical Structural\nPriors Improve Neural Molecular\nRepresentations in Molecule Language Models","url":"https://doi.org/10.1021/acs.jcim.6c02225","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1021%2Facs.jcim.6c02225","date":"2026-11-16T00:00:00Z","timestamp":1794787200,"categories":["Cheminformatics","Design de novo"],"topic_ids":["cheminformatics","denovo"],"keywords":["molecule generation","chemoinformatics","property prediction","molecular representations"],"matched_keywords":["molecule generation","chemoinformatics","property prediction","molecular representations","Molecular","Molecule","chemical"],"matched_tags":["de_novo_design","cheminformatics","ml_task","molecular_descriptor","generic_chem"],"doi":"10.1021/acs.jcim.6c02225","external_id":"f26770a898f942ae759c60a12c91319e2fffb042","pdf_url":null,"code_url":null,"code_host":null,"authors":["Donghyeok Lee","Sangkwon Park","Munyoung Chang","Sungroh Yoon"],"journal":"Journal of Chemical Information\nand Modeling","publisher":null,"impact_factor":null,"abstract":"Classical molecular similarity is central to chemical information systems, yet molecule-language models trained on paired structure–text data do not necessarily organize molecular embeddings according to graded structural relationships. Here, we introduce Structure-Aware Guided Alignment (SAGA), a training-only regularizer that uses Centered Kernel Alignment to align neural molecular representations with fingerprint-based Tanimoto similarity and requires no additional parameters or fingerprints at inference. In MolCA and Atomas, SAGA improves molecule–text retrieval, retrieves closer structural analogues, and raises average performance in scaffold-split property prediction, without clear degradation of molecular description or text-to-molecule generation. These results, together with control experiments that attribute the gains to the chemical content of the prior, show that established chemoinformatics priors can make molecule-language representations more chemically faithful while retaining their multimodal utility.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:c47d44e8f271abb9c92e0dec1eb2326d3e5776f4","kind":"journals","source":"The Journal of chemical physics","title":"Classical-to-DFT multi-fidelity approach for developing machine learning potentials of aqueous systems","url":"https://doi.org/10.1063/5.0341814","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1063%2F5.0341814","date":"2026-11-16T00:00:00Z","timestamp":1794787200,"categories":["ML Potentials"],"topic_ids":["potentials"],"keywords":["machine learning potentials","neural network potentials","DFT","density functional theory","molecular dynamics"],"matched_keywords":["machine learning potentials","Machine learning interatomic potentials","neural network potentials","neural network","machine learning","DFT","density functional theory","molecular","molecular dynamics"],"matched_tags":["mlip_tools","neural_network","machine_learning","dft","generic_chem","molecular_dynamics"],"doi":"10.1063/5.0341814","external_id":"c47d44e8f271abb9c92e0dec1eb2326d3e5776f4","pdf_url":null,"code_url":null,"code_host":null,"authors":["Md. Abu Bakar Siddique","Kendra Coronado","M. M. Badhon","M. Galib"],"journal":"The Journal of chemical physics","publisher":null,"impact_factor":null,"abstract":"Machine learning interatomic potentials (MLIPs) offer near-ab initio accuracy at a fraction of the cost of conventional electronic structure calculations, but their reliability depends critically on the quality and diversity of training data. Here, we present a classical-to-density functional theory (DFT) multi-fidelity workflow for efficient dataset generation, in which long classical molecular dynamics trajectories are supplemented with selectively DFT-optimized configurations to train neural network potentials. We demonstrate this approach on four systems: bulk water, water at the air-water interface, CaSO4 in bulk water, and CaSO4 at the interface. In all cases, the trained models faithfully reproduce reference DFT properties, validating the workflow across both simple and chemically complex environments. We further probe the extrapolation capabilities of the models, showing that bulk-trained networks can generalize to different temperatures, pressures, and interfacial geometries when datasets include targeted augmented configurations. This multi-fidelity strategy thus enables accurate and transferable MLIPs for complex solvated and interfacial systems, bridging the gap between computational efficiency and first-principles accuracy.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:0145417e1ccbf49141336f289259b343c1132557","kind":"journals","source":"Molecules","title":"Comparative Computational Investigation of Protonated Methamphetamine Inclusion Complexes with β-Cyclodextrin, 4-Sulfocalix[4]arene, and Cucurbit[7]uril","url":"https://doi.org/10.3390/molecules31224151","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.3390%2Fmolecules31224151","date":"2026-11-16T00:00:00Z","timestamp":1794787200,"categories":["Docking & Screening"],"topic_ids":["docking"],"keywords":["Molecular docking","density functional theory","DFT"],"matched_keywords":["Molecular docking","density functional theory","DFT","molecular"],"matched_tags":["docking","dft","generic_chem"],"doi":"10.3390/molecules31224151","external_id":"0145417e1ccbf49141336f289259b343c1132557","pdf_url":null,"code_url":null,"code_host":null,"authors":["P. Toochinda","Luckhana Lawtrakul"],"journal":"Molecules","publisher":null,"impact_factor":null,"abstract":"The molecular recognition of protonated methamphetamine (METH) enantiomers by β-cyclodextrin (β-CD), 4-sulfocalix[4]arene (SC[4]), and cucurbit[7]uril (CB[7]) was investigated using a comparative computational approach. Molecular docking was employed to identify favorable inclusion geometries, followed by density functional theory (DFT) calculations using the B97-D3/def2-SVP level of theory with the SMD water solvation model. Natural bond orbital (NBO), molecular electrostatic potential (MEP), reduced density gradient (RDG), noncovalent interaction (NCI), and independent gradient model based on Hirshfeld partitioning (IGMH) analyses were performed to characterize the host–guest interactions. Thermodynamic properties were evaluated using rigid-rotor harmonic-oscillator (RRHO) and quasi-RRHO treatments, while counterpoise-corrected interaction energies were calculated at the ωB97X-D/def2-TZVP level. The calculated energy differences between R-METH and S-METH complexes reflect the selected optimized configurations and limitations of conformational sampling rather than definitive evidence of equilibrium enantioselectivity, particularly for the achiral hosts SC[4] and CB[7]. Comparison with experimental binding constants indicates that METH exhibits stronger binding with CB[7] and SC[4] than with β-CD, although the reported binding affinities vary among experimental methods. Overall, these findings highlight the influence of host molecular architecture, guest orientation, and thermochemical treatment on the calculated stability and intermolecular interactions of METH inclusion complexes, while emphasizing the distinction between structure-dependent complexation energies and equilibrium thermodynamic binding properties.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:4bf3ec9c71219ffb2ff10317d1c56da5c2e958af","kind":"journals","source":"Current Issues in Molecular Biology","title":"Computational Characterization of Ceftriaxone and Novel Compounds in Opioid Use Disorder: Network Pharmacology, Molecular Docking, and Molecular Dynamics Simulations","url":"https://doi.org/10.3390/cimb48111231","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.3390%2Fcimb48111231","date":"2026-11-16T00:00:00Z","timestamp":1794787200,"categories":["Docking & Screening"],"topic_ids":["docking"],"keywords":["Molecular Docking","AutoDock Vina","Molecular Dynamics"],"matched_keywords":["Molecular Docking","AutoDock Vina","Compounds","Molecular","Molecule","Molecular Dynamics","protein"],"matched_tags":["docking","docking_software","generic_chem","molecular_dynamics","biology"],"doi":"10.3390/cimb48111231","external_id":"4bf3ec9c71219ffb2ff10317d1c56da5c2e958af","pdf_url":null,"code_url":null,"code_host":null,"authors":["Rawan S. Alsikhan","Faizul Azam","M. M. Alshehri","Q. Alqahtani","Hala H. Alrefashy","Sarah T. Binakeel","A. Alasmari","Y. Sari","W. Childers","M. Abou-Gharbia","Thikra Algahtani","F. Alasmari"],"journal":"Current Issues in Molecular Biology","publisher":null,"impact_factor":null,"abstract":"Opioid use disorder (OUD) is a global issue characterized by compulsive opioid use and is linked to significant morbidity and mortality. Ceftriaxone has been investigated for its potential neuroprotective and immunomodulatory effects; however, its mechanisms in opioid dependence remain unclear. This study employed a network pharmacological approach to explore the potential molecular target of ceftriaxone and related signaling pathways in OUD. Targets were identified based on a comparative toxicogenomics database, SwissTargetPrediction, complemented by PharmMapper for ceftriaxone and GeneCards and Gene Expression Omnibus for OUD. Overlapping targets were identified using Venn diagrams and analyzed via the STRING platform to construct a protein–protein interaction network. Biological insights were sourced from the DAVID database, and Gene Ontology and KEGG pathway enrichment evaluated the biological significance of the target genes using Metascape. Molecular docking was performed using AutoDock Vina (version 1.1.1) for ceftriaxone, its analogs, and other GLT-1 modulators: MC-100093, MC-350012, MC-350013, and MC-290401. A total of 10 overlapping genes (IL-1β, TNF, ALB, EGFR, ESR1, CDK5, MME, SLC1A2, EPO, and SRC) were identified with multiple biological functions and signaling pathways. Molecular docking predicted that the novel GLT-1 modulator, MC-350013, exhibited favorable docking scores toward several prioritized proteins, particularly ESR1, EGFR, IL-1β, and ALB. Molecule 13 also demonstrated favorable docking scores for the selected proteins. Molecular dynamics simulations characterized the interaction behavior of the novel GLT-1 modulator, MC-350012 and MC-350013. Although ceftriaxone showed stronger hydrogen-bonding interactions, MC-350012 and MC-350013 showed favorable predicted interaction profiles with IL-1β and ESR1. Collectively, these computational findings may suggest MC-350012 and MC-350013 as candidates for further experimental investigation based on their predicted interactions with key OUD-related targets.","source_metadata":{"source":"semantic_scholar"}},{"id":"preprints:10.26434/chemrxiv.15012270/v1","kind":"preprints","source":"ChemRxiv","title":"Computational Drug Discovery of Natural Product Compounds for Tyrosinase Inhibition","url":"https://doi.org/10.26434/chemrxiv.15012270/v1","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.26434%2Fchemrxiv.15012270%2Fv1","date":"2026-11-16T00:00:00Z","timestamp":1794787200,"categories":["Cheminformatics","Property Prediction","Docking & Screening","ADMET & Safety"],"topic_ids":["cheminformatics","property","docking","admet"],"keywords":["ChEMBL","molecular docking","virtual screening","QSAR","gradient boosting","enzyme","bioactivity"],"matched_keywords":["ChEMBL","molecular docking","virtual screening","QSAR","gradient boosting","enzyme","Drug","Compounds","compound","chemical","molecular","bioactivity"],"matched_tags":["chem_benchmarks","docking","virtual_screening","qsar_qspr","ensemble","biology","generic_chem","bioactivity"],"doi":"10.26434/chemrxiv.15012270/v1","external_id":"10.26434/chemrxiv.15012270/v1","pdf_url":null,"code_url":null,"code_host":null,"authors":["Anand Sharma"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"Plants produce chemically diverse compounds for defence and survival, many of which exhibit biological activity and offer promising starting points for drug discovery. This study thus investigated natural compounds as potential inhibitors of tyrosinase, a key enzyme in melanin production and an established target in the treatment of hyperpigmentation. A quantitative structure–activity relationship (QSAR) machine-learning model was developed using ChEMBL tyrosinase bioactivity data and property-matched decoys. Ten models, including decision trees, random forests and gradient boosting, were evaluated; logistic regression was selected for virtual screening. To make the model generalise to unseen chemistry, the model was tested on compounds that were required to have Tanimoto similarity of 0.40 or less to every training compound and to share no Bemis–Murcko scaffold with them. The model achieved an AUC of 0.82 (95% confidence interval [CI], 0.78–0.86), compared with 0.56 using physicochemical properties alone. The frozen model was then used to screen 404,570 natural products from COCONUT (COlleCtion of Open Natural prodUcTs) and IMPPAT (Indian Medicinal Plants, Phytochemistry And Therapeutics). Of the top 1% of predictions, 1,694 compounds were both structurally novel and represented new scaffolds. Following computational filters for predicted skin permeability and potentially reactive chemical groups indicating skin irritation, the 100 highest-scoring candidates were selected for molecular docking. Docking reproduced the crystallographic tropolone binding pose with a root-mean-square deviation (RMSD) of 1.20 Å, while the most predictive docking score achieved an AUC of 0.64 (95% CI, 0.61–0.66). Among the ten highest-ranked candidates, two were associated with Anisomeles indica (Indian catmint) and two with Phaseolus lunatus (lima beans). These candidates warrant experimental validation and further investigation for potential dermatological applications.","source_metadata":{"date_field":"posted","keywords":[]}},{"id":"preprints:10.26434/chemrxiv.15012252/v1","kind":"preprints","source":"ChemRxiv","title":"Computational Studies on the Stability of BINOL-CPA/Imine Complexes","url":"https://doi.org/10.26434/chemrxiv.15012252/v1","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.26434%2Fchemrxiv.15012252%2Fv1","date":"2026-11-16T00:00:00Z","timestamp":1794787200,"categories":["Free Energy & MD"],"topic_ids":["freeenergy"],"keywords":["thermodynamic integration","density functional theory","force fields","molecular dynamics"],"matched_keywords":["thermodynamic integration","density functional theory","force fields","chemical","molecular","molecular dynamics"],"matched_tags":["free_energy_calcs","dft","force_fields","generic_chem","molecular_dynamics"],"doi":"10.26434/chemrxiv.15012252/v1","external_id":"10.26434/chemrxiv.15012252/v1","pdf_url":null,"code_url":null,"code_host":null,"authors":["Emily Groß","Dominik Horinek"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"Chiral BINOL-derived phosphoric acids (CPAs) are widely used Brønsted acid catalysts in enantioselective synthesis, where the relative stabilities of experimentally accessible binary CPA/imine complexes have been linked to stereoselectivity, yet previous quantum-chemical studies using implicit solvation showed substantial deviations from their observed E/Z ratios. Here, we combine density functional theory with explicit-solvent molecular dynamics simulations in dichloromethane, using quantum-chemically derived force fields together with thermodynamic integration and umbrella sampling to determine solvation and complexation free energies. For the investigated systems, the calculated free-energy differences reproduce the experimental E/Z ratios with good accuracy and quantify the relative contributions of solvation and complexation, demonstrating the value of explicit-solvent free-energy simulations for describing CPA/imine equilibria. The method also enables conformational analysis, revealing that the Z isomers access a broader range of relative CPA–imine configurations than the E isomers, suggesting an entropic contribution to their stabilization.","source_metadata":{"date_field":"posted","keywords":[]}},{"id":"journals:3be16988b2961fae476ab126e3ef8bf81c136fb9","kind":"journals","source":"RSC Advances","title":"Design and synthesis of some benzothiazole hybrids as VEGFR-2-targeting anti-prostate cancer agents and their computational studies","url":"https://doi.org/10.1039/d6ra07735h","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1039%2Fd6ra07735h","date":"2026-11-16T00:00:00Z","timestamp":1794787200,"categories":["Docking & Screening"],"topic_ids":["docking"],"keywords":["Molecular docking","DFT","inhibition constants","IC50"],"matched_keywords":["Molecular docking","DFT","inhibition constants","compounds","compound","Molecular","IC50"],"matched_tags":["docking","dft","inhibition_constant","generic_chem","bioactivity"],"doi":"10.1039/d6ra07735h","external_id":"3be16988b2961fae476ab126e3ef8bf81c136fb9","pdf_url":null,"code_url":null,"code_host":null,"authors":["Ramana Gujjula","S. Marupati","B. Patel","R. Talaviya","G. N. Reddy","K. Kapadiya"],"journal":"RSC Advances","publisher":null,"impact_factor":null,"abstract":"Herein, we synthesized some benzothiazole–thiazolidine-2,4-dione–isoxazole hybrids (7a–7n) via Knoevenagel condensation and Cu(i)-catalyzed one-pot regioselective 1,3-dipolar cycloaddition as key approaches. Later, an MTT assay was used to assess the anti-prostate cancer efficacy of the hybrids (7a–7n) against human cancer cell lines (PC3 and DU-145). Among them, compounds 7f–7i surpassed 5-fluorouracil (5-FU) (IC50 = 37.8 µM) against DU-145, with IC50 values in the range of 21.2–34.2 µM. Additionally, compound 7f showed comparable VEGFR-2 inhibition (IC50 = 32.3 nM) to sorafenib (IC50 = 31.5 nM), while compound 7g showed promising VEGFR-2 inhibition (IC50 = 36.6 nM) compared to sorafenib. Molecular docking against VEGFR-2 (PDB ID: 3VHE) revealed that compounds 7f–7i showed better binding energies and inhibition constants than sorafenib. Interestingly, the binding energies and inhibition constants of 7f and 7g were found to be consistent with the VEGFR-2 inhibition data. Finally, the DFT investigation of 7f revealed a conjugated donor–π–acceptor system with good charge transfer properties. The negative electrostatic potential of carbonyl oxygen suggests possible interactions with VEGFR-2.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:10.1002/cmdc.70538","kind":"journals","source":"ChemMedChem","title":"Design, Synthesis and Anti‐Proliferative Assessment of 4‐Aminoquinoline‐ and 4‐Aminoquinoline‐Benzoxaborole Conjugates","url":"https://doi.org/10.1002/cmdc.70538","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1002%2Fcmdc.70538","date":"2026-11-16T00:00:00+00:00","timestamp":1794787200,"categories":["Docking & Screening"],"topic_ids":["docking"],"keywords":["molecular docking","receptor"],"matched_keywords":["molecular docking","compound","compounds","molecular","ligand","receptor"],"matched_tags":["docking","generic_chem","biology"],"doi":"10.1002/cmdc.70538","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Anu Saini","Sumit Kumar","Mandeep Kaur","Sourav Taru Saha","Raghu Raj","Vipan Kumar"],"journal":"ChemMedChem","publisher":"Wiley","impact_factor":3.4,"abstract":"A series of 4‐aminoquinoline analogues and 4‐aminoquinoline‐benzoxaborole conjugates were synthesized and evaluated for their antiproliferative activities against two distinct breast cancer cell lines: MCF‐7 (estrogen receptor‐positive, ER+) and MDA‐MB‐231 (triple‐negative). Preliminary dose‐response assays revealed compound 5e , a 4‐aminoquinoline–benzoxaborole conjugate, as the most active compound against MCF‐7 cells. Structure–activity relationship (SAR) analysis indicated that the cytotoxic potency of the series increased with the length of the alkyl spacer connecting the pharmacophoric moieties. To assess selectivity, HEK‐293 cells were employed as a non‐cancerous control line, and most compounds exhibited minimal cytotoxicity, except compound 5d , which showed moderate effects. To gain preliminary structural insights, exploratory molecular docking studies were performed against the ligand‐binding domain of estrogen receptor alpha (ERα). This integrated experimental‐computational approach highlights the identification of promising preliminary hits. Overall, the findings show that 4‐aminoquinoline–benzoxaborole conjugates exhibit antiproliferative activity against breast cancer cells, with greater potency towards MCF‐7 cells.","source_metadata":{"short_name":"ChemMedChem","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"journals:6a5e66bc9467a2de6ba9e911e494dfa90ea458b0","kind":"journals","source":"Frontiers in Cell and Developmental Biology","title":"Dioscin attenuates OGD/R-induced microglial inflammation: involvement of SIRT1-Associated NADPH redox homeostasis and FSP1/CoQ antiferroptotic defense","url":"https://doi.org/10.3389/fcell.2026.1978045","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.3389%2Ffcell.2026.1978045","date":"2026-11-16T00:00:00Z","timestamp":1794787200,"categories":[],"topic_ids":[],"keywords":["Molecular docking","molecular dynamics"],"matched_keywords":["Molecular docking","Molecular","molecular dynamics","protein"],"matched_tags":["docking","generic_chem","molecular_dynamics","biology"],"doi":"10.3389/fcell.2026.1978045","external_id":"6a5e66bc9467a2de6ba9e911e494dfa90ea458b0","pdf_url":null,"code_url":null,"code_host":null,"authors":["Dan-Dan Liu","Chang Chai","Yi-Die Bao","Jian Li","Jia-Chi Zhang","Wen-Xiang Li","Xiao-Dan Liu"],"journal":"Frontiers in Cell and Developmental Biology","publisher":null,"impact_factor":null,"abstract":"Dioscin exhibits anti-inflammatory, antioxidant, and neuroprotective activities. However, its regulatory effects on microglial ferroptosis and inflammatory responses, as well as the underlying mechanisms, remain unclear. This study aimed to investigate whether Dioscin attenuates oxygen–glucose deprivation/reoxygenation (OGD/R)-induced ferroptosis and inflammatory responses in microglia through a SIRT1-dependent mechanism involving the NADPH-associated FSP1/CoQ defense system. Using OGD/R-treated BV2 microglial cells as an in vitro model, we found that Dioscin dose-dependently ameliorated the OGD/R-induced reduction in cell viability and attenuated iron accumulation, lipid peroxidation, and proinflammatory cytokine release. Mechanistically, Dioscin increased SIRT1 expression and deacetylase activity, reduced the NADP + /NADPH ratio, and elevated membrane-associated FSP1 levels and the CoQ10H 2 /CoQ10 ratio. It also upregulated GPX4 expression while downregulating ACSL4 expression. Both the SIRT1 inhibitor EX-527 and siRNA-mediated knockdown of FSP1 attenuated the beneficial effects of Dioscin on oxidative stress- and ferroptosis-related markers (SOD, GSH, MDA, and Fe 2+ ) and proinflammatory cytokine levels (IL-1β, TNF-α, and IL-6). Further experiments showed that SIRT1 knockdown significantly increased the NADP + /NADPH ratio while reducing membrane-associated FSP1 levels and the CoQ10H 2 /CoQ10 ratio. Supplementation with exogenous NADPH significantly restored the NADP + /NADPH ratio and partially recovered membrane-associated FSP1 levels and CoQ redox status, but failed to restore total FSP1 protein levels. Subcellular fractionation analysis further revealed that Dioscin predominantly increased FSP1 levels in the membrane fraction. Molecular docking and molecular dynamics simulations suggested a potentially stable interaction between Dioscin and SIRT1. Collectively, these findings demonstrate that Dioscin attenuates OGD/R-induced ferroptosis-related injury and inflammatory responses in BV2 microglial cells. Its protective effects are closely associated with SIRT1-related NADPH redox homeostasis and membrane-associated FSP1/CoQ antioxidant defense, accompanied by changes in GPX4 and ACSL4 expression. The present study provides in vitro experimental evidence to support further investigation into the potential protective effects of Dioscin against cerebral ischemia–reperfusion-related cellular injury.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:0ce7757aaf81f40306d5a9c108e5e682e31863a7","kind":"journals","source":"BioNanoScience","title":"Electronic Structure Modulation of MoS2 Nanosheets by Ni-Doping for Antibacterial Activities: Experimental Investigations Supported by DFT and In Silico Molecular Docking Analysis","url":"https://doi.org/10.1007/s12668-026-02976-w","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1007%2Fs12668-026-02976-w","date":"2026-11-16T00:00:00Z","timestamp":1794787200,"categories":["Docking & Screening"],"topic_ids":["docking"],"keywords":["Molecular Docking","DFT"],"matched_keywords":["Molecular Docking","DFT","Molecular"],"matched_tags":["docking","dft","generic_chem"],"doi":"10.1007/s12668-026-02976-w","external_id":"0ce7757aaf81f40306d5a9c108e5e682e31863a7","pdf_url":null,"code_url":null,"code_host":null,"authors":["Rajwinder Singh","Ruchika","Vikash Kushwaha","Sushil Kumar","J. Shahi","D. Mehta","Sanjeev Kumar"],"journal":"BioNanoScience","publisher":null,"impact_factor":null,"abstract":"","source_metadata":{"source":"semantic_scholar"}},{"id":"preprints:10.26434/chemrxiv.15011908/v3","kind":"preprints","source":"ChemRxiv","title":"End-to-End Learning from Experimental Data: Integrating the PC-SAFT Equation of State into Neural Network Training for Improved Polymer Parameter Estimation","url":"https://doi.org/10.26434/chemrxiv.15011908/v3","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.26434%2Fchemrxiv.15011908%2Fv3","date":"2026-11-16T00:00:00Z","timestamp":1794787200,"categories":["Cheminformatics","Property Prediction"],"topic_ids":["cheminformatics","property"],"keywords":["ECFP","graph neural network","GNN","molecular representation"],"matched_keywords":["ECFP","Neural Network","graph neural network","GNN","machine learning","ML","deep learning","molecular representation","chemical","molecular","molecules"],"matched_tags":["molecular_fingerprints","neural_network","gnn","machine_learning","molecular_descriptor","generic_chem"],"doi":"10.26434/chemrxiv.15011908/v3","external_id":"10.26434/chemrxiv.15011908/v3","pdf_url":null,"code_url":null,"code_host":null,"authors":["Hasan Ismaeel","Eleonora Ricci","Panagiotis Krokidas","Vassilis Gkatsis","George Giannakopoulos","Maria Grazia De Angelis"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"Accurate and computationally efficient thermodynamic models are essential for the design, optimization, and operation of many (bio)chemical processes involving polymer solutions, blends, and membranes. In this work, we combine machine learning (ML) with the PC-SAFT equation of state (EoS) to predict the non-associating pure component parameters for polymers, enabling the rapid evaluation of the pressure–volume–temperature (PVT) behavior of polymer melts and rubbers. Using a dataset of more than 5,400 PVT measurements for 78 chemically diverse polymers, we systematically evaluate the effects of EoS integration in the training process, the molecular representation, neural network architecture, and transfer learning. A baseline approach trained on pre-fitted PC-SAFT parameters is compared with a physics-informed approach that embeds the EoS in the training loop, enabling end-to-end training directly on experimental PVT data. Performance is assessed using 6-fold cross-validation. For a feed-forward neural network (FNN) using extended-connectivity fingerprints (ECFPs), embedding the EoS reduces the average mean absolute percentage error (MAPE) from 8.67% to 5.57%. Replacing the FNN and ECFP representation with a graph neural network (GNN) using cyclic polymer graphs further reduces the average MAPE to 2.58%. Pre-training the models on PC-SAFT parameters for small molecules, followed by fine-tuning on polymer PVT data, yields additional improvements, reducing the average MAPE to 5.06% for the FNN and 2.15% for the GNN. These results demonstrate the benefits of combining an embedded EoS with geometric deep learning and transfer learning for accurate polymer PVT predictions.","source_metadata":{"date_field":"posted","keywords":[]}},{"id":"journals:10.1021/jacs.6c11940","kind":"journals","source":"Journal of the American Chemical Society","title":"Expanding Chemical Space Enables Data–Driven\nStructure–Activity Relationships in Cage Catalysis","url":"https://doi.org/10.1021/jacs.6c11940","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1021%2Fjacs.6c11940","date":"2026-11-16T00:00:00+00:00","timestamp":1794787200,"categories":[],"topic_ids":[],"keywords":["random forest","DFT","molecular features","molecular descriptors"],"matched_keywords":["random forest","DFT","molecular features","molecular descriptors","Chemical","molecular"],"matched_tags":["ensemble","dft","molecular_descriptor","generic_chem"],"doi":"10.1021/jacs.6c11940","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Laura E. Bickerton","Tomasz K. Piskorz","Benjamin Norris","Jun Zhou","Yumeng Liu","Jianzhu Wang","Gary S. Nichol","James E. M. Lewis","Fernanda Duarte","Paul J. Lusby"],"journal":"Journal of the American Chemical Society","publisher":"ACS","impact_factor":15.7,"abstract":"Data-driven approaches are increasingly popular in catalysis, supplementing and even replacing traditional strategies that relied on empirical optimization. However, their application in the field of bioinspired cage catalysis remains largely unexplored, primarily due to the limited chemical space of known catalytically active cages. As a result, mechanistic insights are often obtained using computationally intensive DFT calculations on a handful of structurally similar cages. As chemical systems grow in complexity, this approach becomes impractical. Herein, we report a framework that significantly expedites the exploration of a much wider structural landscape while offering computational insights. This is made possible by a novel in situ experimental procedure─which circumvents the traditional time-consuming process of prepare, isolate, characterize, and test─coupled with data-driven methods to identify molecular features that govern activity. Using this approach, we experimentally explore 30 unique PdnL2n cages (n = 2–4) that mediate a nitro-aldol reaction with activities spanning over 4 orders of magnitude. Using this data set, a random forest model trained on molecular descriptors was developed to predict ΔG‡ (MAE of 0.4 kcal mol–1 and R2 = 0.9). SHapley Additive exPlanations (SHAP) analysis highlighted geometric preorganization and electrostatic interactions within the cage as key determinants of catalytic efficiency. Moreover, DFT calculations on a selection of cages further validated these insights, indicating that meaningful mechanistic understanding can be achieved at significantly reduced computational cost. We anticipate that this integrated, screening, data-driven, mechanistic strategy will provide a blueprint workflow for understanding metal–organic cage reactivity and encourage the generation of more comprehensive structural data sets, ultimately enabling the rational design of improved catalytic architectures.","source_metadata":{"short_name":"JACS","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"journals:4fcef1129e559d50b9d249472b684a9ee145b612","kind":"journals","source":"Journal of\nthe American Chemical Society","title":"Expanding Chemical Space Enables Data–Driven\nStructure–Activity Relationships in Cage Catalysis","url":"https://doi.org/10.1021/jacs.6c11940","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1021%2Fjacs.6c11940","date":"2026-11-16T00:00:00Z","timestamp":1794787200,"categories":[],"topic_ids":[],"keywords":["random forest","DFT","molecular features","molecular descriptors"],"matched_keywords":["random forest","DFT","molecular features","molecular descriptors","Chemical","molecular"],"matched_tags":["ensemble","dft","molecular_descriptor","generic_chem"],"doi":"10.1021/jacs.6c11940","external_id":"4fcef1129e559d50b9d249472b684a9ee145b612","pdf_url":null,"code_url":null,"code_host":null,"authors":["L. Bickerton","T. Piskorz","Benjamin Norris","Jun Zhou","Yu-Meng Liu","Jian-Zhu Wang","G. Nichol","James E. M. Lewis","Fernanda Duarte","P. J. Lusby"],"journal":"Journal of\nthe American Chemical Society","publisher":null,"impact_factor":null,"abstract":"Data-driven approaches are increasingly popular in catalysis, supplementing and even replacing traditional strategies that relied on empirical optimization. However, their application in the field of bioinspired cage catalysis remains largely unexplored, primarily due to the limited chemical space of known catalytically active cages. As a result, mechanistic insights are often obtained using computationally intensive DFT calculations on a handful of structurally similar cages. As chemical systems grow in complexity, this approach becomes impractical. Herein, we report a framework that significantly expedites the exploration of a much wider structural landscape while offering computational insights. This is made possible by a novel in situ experimental procedure─which circumvents the traditional time-consuming process of prepare, isolate, characterize, and test─coupled with data-driven methods to identify molecular features that govern activity. Using this approach, we experimentally explore 30 unique PdnL2n cages (n = 2–4) that mediate a nitro-aldol reaction with activities spanning over 4 orders of magnitude. Using this data set, a random forest model trained on molecular descriptors was developed to predict ΔG‡ (MAE of 0.4 kcal mol–1 and R2 = 0.9). SHapley Additive exPlanations (SHAP) analysis highlighted geometric preorganization and electrostatic interactions within the cage as key determinants of catalytic efficiency. Moreover, DFT calculations on a selection of cages further validated these insights, indicating that meaningful mechanistic understanding can be achieved at significantly reduced computational cost. We anticipate that this integrated, screening, data-driven, mechanistic strategy will provide a blueprint workflow for understanding metal–organic cage reactivity and encourage the generation of more comprehensive structural data sets, ultimately enabling the rational design of improved catalytic architectures.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:0191fc6cb992e9c4e223cdb398d5093c295844f3","kind":"journals","source":"Talanta","title":"Fluorescent/colorimetric dual-mode sensor based on bifunctional Fe-N-C rhombic dodecahedron combined with a triple-helix molecular structure for the detection of acrylamide in foods","url":"https://doi.org/10.1016/j.talanta.2026.130965","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1016%2Fj.talanta.2026.130965","date":"2026-11-16T00:00:00Z","timestamp":1794787200,"categories":["Docking & Screening"],"topic_ids":["docking"],"keywords":["molecular docking"],"matched_keywords":["molecular docking","molecular"],"matched_tags":["docking","generic_chem"],"doi":"10.1016/j.talanta.2026.130965","external_id":"0191fc6cb992e9c4e223cdb398d5093c295844f3","pdf_url":null,"code_url":null,"code_host":null,"authors":["Jia-Yu Niu","Meng-Hui Wang","Feng-Juan Zhao","Zhi-Guang Suo","Yong Liu","Min Wei"],"journal":"Talanta","publisher":null,"impact_factor":null,"abstract":"In this study, a fluorescent/colorimetric dual-mode aptasensor based on the bifunctional Fe-N-C rhombic dodecahedron (Fe-N-C RD) and triple-helix molecular structure (THMS) amplification strategy was developed for the detection of acrylamide (AA) in foods. The optimal APT-1 was selected as the recognition element through molecular docking and experimental validation. Compared to N-C material, Fe-N-C RD exhibited 2.02-fold increase in quenching performance and excellent peroxidase-like activity (lower Km, higher Vmax). The results demonstrated that the fluorescent signal of Fe-N-C RD/THMS strategy was enhanced by 2.08-fold and 1.19-fold compared with those of Fe-N-C RD/T2/CDs@T3 and Fe-N-C RD/CDs@T1, while the colorimetric signal was increased by 1.44-fold and 1.48-fold, respectively. In fluorescent mode, a linear relationship was observed between ΔF and the logarithmic AA concentration (lgCAA) across the range of 0.01-1000 nM, with the linear equation of ΔF = (197.250 ± 4.235)lgCAA + (501.728 ± 7.519) (R2 = 0.996) and the limit of detection (LOD) of 8.49 pM. In colorimetric mode, a linear correlation between ΔA and lgCAA was achieved within the range of 0.1-500 nM, with the linear equation of ΔA = (0.148 ± 0.005)lgCAA + (0.235 ± 0.007) (R2 = 0.993) and the LOD of 35.57 pM. Furthermore, the signal normalization of the dual-mode sensor was achieved through weighted processing, and the linear equation was Yw = (97.345 ± 3.282)lgCAA + (248.055 ± 3.897) (R2 = 0.995) within the range of 0.1-500 nM, with the LOD of 21.99 pM. The sensor was successfully applied to detect AA in actual samples (coffee, potato chips) and quality control samples (potato chips, cookies). The recoveries of potato chips and coffee were in the range of 93.12%-109.29% in fluorescent mode and 91.65%-116.33% in colorimetric mode, with RSD of less than 15% for both samples.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:10.1021/acs.jctc.6c01672","kind":"journals","source":"Journal of Chemical Theory and Computation","title":"Force Calculation\nat the Complete Active Space Configuration\nInteraction Level Based on the Neural Network Quantum State Approach","url":"https://doi.org/10.1021/acs.jctc.6c01672","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1021%2Facs.jctc.6c01672","date":"2026-11-16T00:00:00+00:00","timestamp":1794787200,"categories":[],"topic_ids":[],"keywords":["quantum chemistry"],"matched_keywords":["Neural Network","quantum chemistry","chemical"],"matched_tags":["neural_network","quantum_chemistry","generic_chem"],"doi":"10.1021/acs.jctc.6c01672","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Mingxin Li","Juntao Lai","Yumeng Zhou","Qiang Fu","Honghui Shang","Zhenyu Li","Jinlong Yang"],"journal":"Journal of Chemical Theory and Computation","publisher":"ACS","impact_factor":null,"abstract":"The neural network quantum state (NNQS) approach provides an artificial intelligence-based representation of many-electron wave functions for quantum chemistry. Although a full configuration-space NNQS can achieve high accuracy, its direct use for force evaluation and geometry optimization remains computationally demanding, motivating a reduced-space treatment. Here, we implement analytic interatomic force calculations at the complete active space configuration interaction (CASCI) level based on the highly expressive QiankunNet architecture. By constructing an effective second-quantized Hamiltonian within a prescribed active orbital space and solving the coupled-perturbed Hartree–Fock equations, we obtain analytic forces including orbital response contributions. Benchmark calculations demonstrate the compact representational capability of QiankunNet and its close agreement with conventional CASCI references for both energies and forces, even in strongly correlated systems. We further apply the resulting forces to construct reaction potential energy surfaces for both ethylene torsion and cyclobutadiene isomerization, and to optimize the anthracene geometry for vertical singlet–triplet energy gap calculation. This work offers a practical route toward efficient force-based simulations of multireference chemical systems.","source_metadata":{"short_name":"JCTC","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"journals:a0c142890a17fcf849ac87edd8b0e884ac675832","kind":"journals","source":"Journal of Chemical Theory\nand Computation","title":"Force Calculation\nat the Complete Active Space Configuration\nInteraction Level Based on the Neural Network Quantum State Approach","url":"https://doi.org/10.1021/acs.jctc.6c01672","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1021%2Facs.jctc.6c01672","date":"2026-11-16T00:00:00Z","timestamp":1794787200,"categories":[],"topic_ids":[],"keywords":["quantum chemistry"],"matched_keywords":["Neural Network","quantum chemistry","chemical"],"matched_tags":["neural_network","quantum_chemistry","generic_chem"],"doi":"10.1021/acs.jctc.6c01672","external_id":"a0c142890a17fcf849ac87edd8b0e884ac675832","pdf_url":null,"code_url":null,"code_host":null,"authors":["Ming-Xin Li","Jun-Tao Lai","Yu-Meng Zhou","Qiang Fu","Hong-Hui Shang","Zhen-Yu Li","Jin-Long Yang"],"journal":"Journal of Chemical Theory\nand Computation","publisher":null,"impact_factor":null,"abstract":"The neural network quantum state (NNQS) approach provides an artificial intelligence-based representation of many-electron wave functions for quantum chemistry. Although a full configuration-space NNQS can achieve high accuracy, its direct use for force evaluation and geometry optimization remains computationally demanding, motivating a reduced-space treatment. Here, we implement analytic interatomic force calculations at the complete active space configuration interaction (CASCI) level based on the highly expressive QiankunNet architecture. By constructing an effective second-quantized Hamiltonian within a prescribed active orbital space and solving the coupled-perturbed Hartree–Fock equations, we obtain analytic forces including orbital response contributions. Benchmark calculations demonstrate the compact representational capability of QiankunNet and its close agreement with conventional CASCI references for both energies and forces, even in strongly correlated systems. We further apply the resulting forces to construct reaction potential energy surfaces for both ethylene torsion and cyclobutadiene isomerization, and to optimize the anthracene geometry for vertical singlet–triplet energy gap calculation. This work offers a practical route toward efficient force-based simulations of multireference chemical systems.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:c777493738d314ef5ab42799cf167200efe56d9b","kind":"journals","source":"International Journal of Molecular Sciences","title":"From Endothelial Survival to Lung Protection: Calycosin Attenuates Ischemia–Reperfusion Injury","url":"https://doi.org/10.3390/ijms272210275","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.3390%2Fijms272210275","date":"2026-11-16T00:00:00Z","timestamp":1794787200,"categories":["Docking & Screening"],"topic_ids":["docking"],"keywords":["molecular docking"],"matched_keywords":["molecular docking","molecular","protein"],"matched_tags":["docking","generic_chem","biology"],"doi":"10.3390/ijms272210275","external_id":"c777493738d314ef5ab42799cf167200efe56d9b","pdf_url":null,"code_url":null,"code_host":null,"authors":["Heng Zhao","Zhe Chen","Chi Wang","Yi-Xing Li","Yi-Han Lin","Hong-Yi Wang","Ming Li","Jin-Teng Feng","Guang-Jian Zhang"],"journal":"International Journal of Molecular Sciences","publisher":null,"impact_factor":null,"abstract":"Lung ischemia–reperfusion injury (LIRI) contributes to early graft dysfunction after lung transplantation, yet effective pharmacological interventions remain limited. Calycosin (CA), an isoflavone sourced from Astragali Radix, has shown protective effects in experimental ischemic injury, but its effects on LIRI remain insufficiently characterized. We integrated network analysis and molecular docking with experiments in human umbilical vein endothelial cells subjected to oxygen–glucose deprivation/reoxygenation (OGD/R) and male Sprague–Dawley rats undergoing left pulmonary hilar occlusion and reperfusion. Computational analyses highlighted inflammatory signaling, redox regulation, and cell survival, with CA improving endothelial viability and reducing oxidative stress, apoptosis, and inflammatory cytokine levels. In rats, CA reduced histological injury, edema, inflammatory cell accumulation, and lipid peroxidation while improving arterial oxygenation, superoxide dismutase activity, and tissue ATP content. Protection was accompanied by changes in candidate gene expression and PPARγ, COX-2, BAX, and BCL2 protein abundance. In endothelial cells, the PPARγ antagonist GW9662 attenuated the protective effects of CA and its reduction of NF-κB p65 nuclear accumulation. These findings support CA’s potential as a candidate intervention for LIRI and identify PPARγ as a contributor to its endothelial–protective effects, providing a rationale for further evaluation in lung preservation and transplantation models.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:10.1007/s00894-026-07056-0","kind":"journals","source":"Journal of Molecular Modeling","title":"From QSAR-guided fragment design to 500 ns dynamics: multiscale mechanistic dissection of novel MNK2 inhibitors","url":"https://doi.org/10.1007/s00894-026-07056-0","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1007%2Fs00894-026-07056-0","date":"2026-11-16T00:00:00+00:00","timestamp":1794787200,"categories":["Property Prediction"],"topic_ids":["property"],"keywords":["QSAR"],"matched_keywords":["QSAR"],"matched_tags":["qsar_qspr"],"doi":"10.1007/s00894-026-07056-0","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Jin-Ying Zhao","Jian-Bo Tong","Rong Wang","Ya-Kun Zhang"],"journal":"Journal of Molecular Modeling","publisher":"Springer","impact_factor":null,"abstract":"","source_metadata":{"short_name":"J. Mol. Model.","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"journals:c10f19cb8b907bcb9c195df0845d6847dbc2781d","kind":"journals","source":"The Journal of chemical physics","title":"From spectra to performance: Multimodal machine learning for organic optoelectronic materials discovery","url":"https://doi.org/10.1063/5.0356266","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1063%2F5.0356266","date":"2026-11-16T00:00:00Z","timestamp":1794787200,"categories":[],"topic_ids":[],"keywords":["property prediction","molecular representations"],"matched_keywords":["machine learning","property prediction","molecular representations","molecular"],"matched_tags":["machine_learning","ml_task","molecular_descriptor","generic_chem"],"doi":"10.1063/5.0356266","external_id":"c10f19cb8b907bcb9c195df0845d6847dbc2781d","pdf_url":null,"code_url":null,"code_host":null,"authors":["Yong Yan","Xun Fu","Jie Hao","Hui Zhu","Meng-Na Yu","Xin-Miao Xie","Xue-Feng Guo","Ling-Hai Xie"],"journal":"The Journal of chemical physics","publisher":null,"impact_factor":null,"abstract":"The discovery and rational optimization of optoelectronic materials increasingly rely on heterogeneous characterization data encoding structure, morphology, electronic states, and device properties. Conventional spectral interpretation remains expert-dependent and often treats each modality in isolation, limiting information extraction. Machine learning has improved information extraction from individual spectroscopic modalities, yet single modality approaches often suffer from incomplete structural and physical representations. Multimodal machine learning enables integration of diverse spectroscopic signals into unified representations. In this review, we summarize recent advances in multimodal spectral learning and heterogeneous data fusion for spectra-structure-property relationships in optoelectronic materials. We discuss how the synergistic combination of infrared, Raman, nuclear magnetic resonance, and mass spectrometry improves functional-group recognition, substructure inference, and spectrum-based molecular structure elucidation and raw-spectrum learning for molecular elucidation. We further elaborate on how the fusion of optical spectra, morphology, processing parameters, and molecular representations facilitates quantitative property prediction and materials discovery. Key challenges include simulation-experiment domain gaps, limited multimodal datasets, missing modalities, and insufficient model interpretability. Overall, multimodal spectral learning is emerging as a promising approach for transforming characterization from a descriptive tool into a predictive framework for rational materials design.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:1d871874886a86e08a900e1045a3db0f48a9af2d","kind":"journals","source":"Foods","title":"Gallic Acid Mitigates Zearalenone-Induced Kidney Oxidative Damage and Modulates Gut Microbiota","url":"https://doi.org/10.3390/foods15224113","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.3390%2Ffoods15224113","date":"2026-11-16T00:00:00Z","timestamp":1794787200,"categories":["Docking & Screening"],"topic_ids":["docking"],"keywords":["molecular docking","enzyme","binding affinity"],"matched_keywords":["molecular docking","enzyme","compound","molecular","binding affinity"],"matched_tags":["docking","biology","generic_chem","bioactivity"],"doi":"10.3390/foods15224113","external_id":"1d871874886a86e08a900e1045a3db0f48a9af2d","pdf_url":null,"code_url":null,"code_host":null,"authors":["Hong-Xiao Jiang","Jia-Xi Wang","Si-Yuan Jing","Jia-Qi Fang","Rizeng Meng","Yan Wang","Yi Liu","Na Guo"],"journal":"Foods","publisher":null,"impact_factor":null,"abstract":"Gallic acid (GA), a natural compound with diverse biological activities, exhibits targeted distribution within kidney tissue after ingestion. However, the mechanisms by which GA regulates zearalenone (ZEN)-induced kidney injury remain unclear, particularly concerning the role of gut microbiota and its metabolites in this protective process. This study aimed to elucidate the protective mechanisms of GA against oxidative damage to the kidney and disturbances in gut microbiota caused by ZEN. We evaluated the effects of GA supplementation on ZEN-induced kidney injury in vivo and in vitro, and further investigated its regulatory effects on the gut microbiome and metabolites in ZEN-exposed mice. The results indicated that molecular docking predicted a strong binding affinity between GA and Keap1, and GA activated Nrf2 nuclear translocation, enhanced antioxidant response element (ARE) transcriptional activity, and activated the downstream target genes of the Keap1/Nrf2/ARE signaling pathway. Additionally, GA increased antioxidant enzyme activity, thereby reducing ZEN-induced oxidative damage. Further analysis of gut microbiota and short-chain fatty acid (SCFA) content revealed that adding GA modulated the disturbance of gut microbiota and showed a trend toward restoring the SCFA levels decreased by ZEN, which may be associated with the mitigation of ZEN-induced kidney injury. In conclusion, GA protects mice from ZEN-induced kidney damage, potentially involving the regulation of gut microbiota and SCFAs, providing a theoretical basis for its potential application in mitigating food hazards contaminated with ZEN.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:10.1088/2632-2153/aec2a2","kind":"journals","source":"Machine Learning: Science and Technology","title":"Generative pseudo-force fields for molecular generation","url":"https://doi.org/10.1088/2632-2153/aec2a2","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1088%2F2632-2153%2Faec2a2","date":"2026-11-16T00:00:00+00:00","timestamp":1794787200,"categories":["Design de novo"],"topic_ids":["denovo"],"keywords":["molecular generation","generative models","diffusion models","denoising","force fields"],"matched_keywords":["molecular generation","Generative pseudo-force fields for molecular generation","generative models","machine learning","diffusion models","denoising","force fields","molecular","drug"],"matched_tags":["de_novo_design","generative_drug_design","generative_model","machine_learning","diffusion","force_fields","generic_chem"],"doi":"10.1088/2632-2153/aec2a2","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Stefaan Simon Pierre Hessmann","Khaled Kahouli","Stefan Gugler","Michael Plainer","Frank Noé","Klaus-Robert Mueller","Niklas Wolf Andreas Gebauer"],"journal":"Machine Learning: Science and Technology","publisher":"IOP","impact_factor":6.3,"abstract":"Generating stable molecular conformations typically forces a tradeoff between the physical realism of energy-based relaxation and the sampling efficiency of data-driven generative models. While machine learning force fields (MLFFs) can sample stable conformations by relaxing molecular geometries according to physical forces, they require costly ab-initio training data. Conversely, diffusion models (DMs) learn from equilibrium data alone but are dependent on noise schedules and time-step conditioning. In this work, we propose generative pseudo-force fields (GPFFs) to bridge these paradigms by training an MLFF on a quadratic pseudo-potential energy surface relative to reference equilibrium structures. Because no ab-initio calculations are required for the perturbed geometries, non-equilibrium training data can be generated on the fly by perturbing the equilibria with Gaussian noise. We show that GPFFs constitute a time-step-agnostic variant of variance-exploding DMs: the score comes from the predicted pseudo-forces but because force magnitudes implicitly encode the noise level, no time-step conditioning is needed. Our GPFF can hence be used as a drop-in replacement in standard diffusion sampling (ancestral, Heun) but also facilitates more efficient, adaptive variants and an MLFF inspired direct denoising scheme. Our proposed sampling algorithms support arbitrary structural priors and geometric constraints. On QM9, GPFF reaches 99 % validity with stochastic direct denoising at 256 neural function evaluations (NFE), and about 90 % at just 16 NFE with direct denoising, outperforming a matched diffusion baseline that shares its backbone, training recipe, and noise schedule across all samplers. Combined with custom priors, we showcase the fast and accurate generation process of our method in a molecular editor for a drug design setting.","source_metadata":{"short_name":"MLST","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"journals:096c635ff9bebeadd2ccb3743ac9b17f9c441c4a","kind":"journals","source":"International Journal of Quantum Chemistry","title":"GFN\n Methods as Geometry Engines for Organic Photovoltaic Databases: A Benchmark Against\n BP86\n /def2‐\n SVP","url":"https://doi.org/10.1002/qua.70342","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1002%2Fqua.70342","date":"2026-11-16T00:00:00Z","timestamp":1794787200,"categories":[],"topic_ids":[],"keywords":["DFT","B3LYP"],"matched_keywords":["ML","DFT","B3LYP","Molecular","molecule","molecules","chemical"],"matched_tags":["machine_learning","dft","generic_chem"],"doi":"10.1002/qua.70342","external_id":"096c635ff9bebeadd2ccb3743ac9b17f9c441c4a","pdf_url":null,"code_url":null,"code_host":null,"authors":["Steve Cabrel Teguia Kouam","J.-P. Tchapet Njafa","Raoult Dabou Teukam","Patrick Sorrel Mvoto Kongo","J. Nguenang","S. G. N. Engo"],"journal":"International Journal of Quantum Chemistry","publisher":null,"impact_factor":null,"abstract":"Molecular simulation workflows based on the GFN family of semiempirical methods (GFN1‐xTB, GFN2‐xTB, GFN0‐xTB, and GFN‐FF) are benchmarked as geometry engines against the original BP86/def2‐SVP DFT level of theory of the Harvard Clean Energy Project (CEP) database for high‐throughput organic photovoltaic (OPV) molecule screening. Two representative datasets are evaluated: a 73‐molecule subset of the CEP database of extended ‐conjugated systems and a 56‐molecule subset of the QM9 database of small organic molecules, selected via variance‐based Neyman allocation from stratified and clustered chemical spaces. Structural fidelity is assessed through heavy‐atom RMSD, radius of gyration, equilibrium rotational constants, bond lengths, and bond angles, and topological consistency via the VF2 graph isomorphism algorithm. GFN1‐xTB and GFN2‐xTB achieve the highest structural fidelity for both datasets, with mean hRMSD values of 0.765 and 0.771 Å for CEP and bond length MADs of 0.018 and 0.021 Å, respectively. GFN2‐xTB achieves the best angle accuracy (MAD = 0.70°) among all methods. VF2 topological failure rates are minimal for CEP (1.4% for both GFN1/2‐xTB workflows) but substantial for QM9 with GFN0‐xTB (26.8%) and GFN‐FF (16.1%), indicating that self‐consistent‐charge methods (GFN1/2‐xTB) provide substantially higher topological reliability than non‐self‐consistent methods for chemically diverse molecular spaces. Meanwhile, GFN‐FF provides the largest computational speedup (up to ∼177× relative to BP86/def2‐SVP) with an empirical ‐like apparent scaling over the sampled molecular size range. Sensitivity analyses across solvation environments, basis sets (def2‐SVP vs. def2‐TZVP), and DFT levels of theory (BP86/def2‐SVP vs. B3LYP/6‐31G(2df,p) for QM9) confirm that method rankings are preserved in all cases (rank correlation ρ > 0.98), validating the primary benchmark as robust for relative method comparison. For OPV database generation, GFN1‐xTB and GFN2‐xTB protocols provide the best overall balance of geometric fidelity, topological reliability, and computational throughput for bond‐resolved and graph‐based ML features, while GFN‐FF's workflow is the preferred geometry engine for first‐pass screening of large molecular libraries where topological validation can be applied. These results provide quantitative guidance for the deployment of GFN workflows in high‐throughput computational pipelines for ML‐oriented organic semiconductor discovery.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:10.1002/wcms.70092","kind":"journals","source":"WIREs Computational Molecular Science","title":"Graph Representation Learning, Geometric Deep Learning, and Foundation Models for Biological Network Analysis","url":"https://doi.org/10.1002/wcms.70092","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1002%2Fwcms.70092","date":"2026-11-16T00:00:00+00:00","timestamp":1794787200,"categories":[],"topic_ids":[],"keywords":["transformers","equivariant","Graph neural networks","GNNs"],"matched_keywords":["transformers","equivariant","neural networks","Graph neural networks","GNNs","Deep Learning","machine learning","molecular","protein"],"matched_tags":["transformer","equivariance","neural_network","gnn","machine_learning","generic_chem","biology"],"doi":"10.1002/wcms.70092","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Gleb Svinin","Rebecca Ting Jiin Loo","Enrico Glaab"],"journal":"WIREs Computational Molecular Science","publisher":"Wiley","impact_factor":null,"abstract":"Biological systems are inherently relational. Genes, proteins, metabolites, and cells interact through complex networks that encode the molecular logic of development, homeostasis, and disease. However, conventional machine learning methods that operate on tabular or sequence‐based representations are not well‐suited to capture this graph‐structured organization. Graph neural networks (GNNs) and their modern descendants, such as graph transformers, equivariant architectures, and foundation models, provide a framework for learning directly from biological interaction networks. This review provides a broad, methodologically focused overview of graph representation learning and geometric deep learning approaches applied to biological network analysis. It covers graph convolutional networks, graph attention networks, message‐passing neural networks, graph transformers, equivariant networks for three‐dimensional molecular structures, and large‐scale pre‐trained foundation models. We introduce a taxonomy that distinguishes graph‐native, knowledge‐graph, graph‐informed, and non‐graph biological foundation models, and we organize the discussion of architectures around the design decisions that matter for biological networks, such as heterophily, edge direction and sign, edge uncertainty, and oversquashing. We evaluate how these architectures can be applied to key network biology tasks, such as gene regulatory network inference, protein–protein interaction prediction, multi‐omics network integration, and disease gene prioritization, and we compare the problem formulations underlying these tasks. We also compare representational strategies, such as heterogeneous graphs and hypergraphs, to capture the multilayer complexity of biological systems. Benchmarking resources, open software tools, and key datasets used across these tasks are cataloged. We discuss challenges, such as data sparsity, network incompleteness, uncertainty in graph construction, data leakage, negative sampling, biological interpretability, and limited ground‐truth labels, and we provide practical guidance for method selection and validation. This review is intended for computational biologists, bioinformaticians, and machine learning researchers who are seeking an overview and introduction to this field with an emphasis on its applicability to disease‐relevant omics data analysis.","source_metadata":{"short_name":"WIREs Comput. Mol. Sci.","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"journals:43c81073c8a7d18aa6e427598a5097fc28158d64","kind":"journals","source":"Nanoscale","title":"Graph-based learning of nanopore translocation event signals","url":"https://doi.org/10.1039/d6nr01686c","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1039%2Fd6nr01686c","date":"2026-11-16T00:00:00Z","timestamp":1794787200,"categories":[],"topic_ids":[],"keywords":["graph neural networks"],"matched_keywords":["neural networks","graph neural networks","molecule","molecular"],"matched_tags":["neural_network","gnn","generic_chem"],"doi":"10.1039/d6nr01686c","external_id":"43c81073c8a7d18aa6e427598a5097fc28158d64","pdf_url":null,"code_url":null,"code_host":null,"authors":["Mayukh Kansari","T. Ensslen","J. Behrends","M. Fyta"],"journal":"Nanoscale","publisher":null,"impact_factor":null,"abstract":"Nanopores enable single-molecule analysis by measuring current signals through nanoscale pores in either biological or solid-state membranes. Accurate detection of analyte fingerprints within the pore environment is essential for reading-out the analyte type and sequence. Towards these aims, we develop a framework for robust, semi-supervised detection of molecular nanopore events on the basis of a graph representation of the measured signals. To this end, using experimental signals from DNA translocation through an aerolysin pore, we build a graph-based two-stage workflow employing convolutional and graph neural networks that first perform fast screening followed by deep validation of these nanopore events. The learned model is able to select possible molecular signatures (the current blockades) in the full signal, denoise, validate, and reconstruct them, and predict the morphology of unseen molecular events within the range of the available data. The trained model can efficiently predict the correct event morphology for the same analyte within a 2.4-fold range of transmembrane voltage values not included in the training. The developed partially supervised workflow, linking unsupervised graph representation-learning to raw-signal inference is - at its current state - a computational proof of concept that proposes a direction for the development of modular and generalizable event detection workflows, provided that they are trained on a huge amount of different nanopore experiments.","source_metadata":{"source":"semantic_scholar"}},{"id":"preprints:10.26434/chemrxiv.15011594/v2","kind":"preprints","source":"ChemRxiv","title":"How Chemical Evidence Shapes Mechanistic Reasoning and Prediction by Large Language Models: Fine-Grained Inference in Nucleoside-Derived Supramolecular Gelation","url":"https://doi.org/10.26434/chemrxiv.15011594/v2","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.26434%2Fchemrxiv.15011594%2Fv2","date":"2026-11-16T00:00:00Z","timestamp":1794787200,"categories":["LLMs & Agents"],"topic_ids":["llm"],"keywords":["LLMs","LLM"],"matched_keywords":["LLMs","LLM","Chemical"],"matched_tags":["LLMs","generic_chem"],"doi":"10.26434/chemrxiv.15011594/v2","external_id":"10.26434/chemrxiv.15011594/v2","pdf_url":null,"code_url":null,"code_host":null,"authors":["Ying Wang","Xi Yu"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"Large language models (LLMs) offer a distinctive capability for chemical problem solving by integrating chemical knowledge and experimental evidence through reasoning expressed in natural language. Yet how chemical evidence reshapes their predictions and expressed rationales is not self-evident and has received limited examination, particularly for challenging chemical problems. Here, using nucleoside-derived supramolecular gelation as a testbed, we systematically varied the evidence provided to LLMs and quantitatively evaluated their predictions and the mechanistic alignment of generated rationales. Without additional evidence, correct predictions were associated with greater alignment to literature-supported mechanisms than incorrect ones, suggesting that LLMs can draw on intrinsic chemical knowledge despite rationales remain largely general. Structurally similar gelation precedents, thought improved prediction accuracy significantly, reduced this distinction by anchoring rationales to insufficiently discriminative structural features, indicating that similarity-based evidence does not necessarily improve chemical reasoning and that prediction accuracy alone is insufficient to assess LLM performance in chemical tasks. In contrast, condition-constrained references and mechanistic information more consistently directed rationales toward chemically discriminative features, even when prediction accuracy did not improve in tandem. Together, our study highlights the importance of task-specific evidence design and quantitative rationale evaluation for reliable chemical problem solving by LLMs.","source_metadata":{"date_field":"posted","keywords":[]}},{"id":"journals:10.1038/s41467-026-79240-y","kind":"journals","source":"Nature Communications","title":"Improving molecular property prediction with score-based diffusion models","url":"https://doi.org/10.1038/s41467-026-79240-y","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1038%2Fs41467-026-79240-y","date":"2026-11-16T00:00:00+00:00","timestamp":1794787200,"categories":["Property Prediction"],"topic_ids":["property"],"keywords":["graph neural networks","property prediction","score based","diffusion models","activity cliffs","molecular property"],"matched_keywords":["neural networks","graph neural networks","property prediction","score-based","diffusion models","activity cliffs","molecular","chemical","molecule","molecular property"],"matched_tags":["neural_network","gnn","ml_task","diffusion","structure_activity_landscape","generic_chem","bioactivity"],"doi":"10.1038/s41467-026-79240-y","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Julien Horwood","Dino Oglic","José Miguel Hernández-Lobato"],"journal":"Nature Communications","publisher":"Nature","impact_factor":15.7,"abstract":"We explore diffusion-based training as a flexible approach to molecular property prediction in settings shaped by chemical, biophysical, and latent heterogeneity. To address this, we build on score-based generative diffusion models and introduce a conditional diffusion module for predictive tasks, equipped with a joint training and sampling guidance scheme and an optional gated residual mechanism that regulates feature- vs target-driven information flow across noise levels. The module acts as a plug-in learning block for existing molecular encoders, making it complementary to advances in representation learning. Across diverse feature extractors and graph neural networks, diffusion-based training improves over conventional maximum-likelihood objectives on small-molecule regression benchmarks and matches or exceeds strong baselines on multitask regression and classification datasets. It also produces well-calibrated predictive uncertainty as measured by prediction interval coverage, enables confidence-guided ranking, and performs well in challenging settings including activity cliffs and class-imbalanced multitask prediction. Overall, these results position diffusion as a promising representation-agnostic approach for effective molecular property prediction.","source_metadata":{"short_name":"Nat. Commun.","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"journals:da989673da87c878525d382b22468411958e95c2","kind":"journals","source":"Current Issues in Molecular Biology","title":"In Silico Identification of a Potential Multi-Target Natural Compound Against Oral Cancer","url":"https://doi.org/10.3390/cimb48111226","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.3390%2Fcimb48111226","date":"2026-11-16T00:00:00Z","timestamp":1794787200,"categories":["Docking & Screening","ADMET & Safety"],"topic_ids":["docking","admet"],"keywords":["drug likeness","molecular docking","DFT","ADMET","MD simulations","Molecular dynamics","receptor"],"matched_keywords":["drug-likeness","molecular docking","DFT","ADMET","MD simulations","Molecular dynamics","Compound","compounds","ligand","molecular","drug","receptor"],"matched_tags":["drug_likeness","docking","dft","admet","molecular_dynamics","generic_chem","biology"],"doi":"10.3390/cimb48111226","external_id":"da989673da87c878525d382b22468411958e95c2","pdf_url":null,"code_url":null,"code_host":null,"authors":["N. Alhussain","A. Alzubaidi","Muhammad Mudassar Saleem","M. Naqash","Muhammad Ishfaq","C. Onwuka","Imran Khalid","S. Sainudeen","A. Salih","Mohasin A. Khader"],"journal":"Current Issues in Molecular Biology","publisher":null,"impact_factor":null,"abstract":"Oral cancer remains a significant global health problem due to delayed diagnosis and limited treatment effectiveness. This study employed an integrated in silico approach to identify potential natural compounds from the family Ganodermataceae, targeting chemokine receptors CXCR3, CXCR4 and CXCR7. A library of 1300 mushroom-derived compounds was screened using a ligand-based pharmacophore. The best hits were subjected to molecular docking. The computational candidate, (+)-Applanatumol D, was further evaluated using ADMET, DFT, 200 ns MD simulations and MM-GBSA. Molecular docking predicted interactions of (+)-Applanatumol D with all three receptors, ranging from −8.0 to −8.5 kcal/mol, with these values interpreted as relative computational docking scores rather than experimentally determined binding affinities. Molecular dynamics simulations showed receptor-dependent structural deviations, with greater deviations observed for the CXCR7 complex. ADMET predictions indicated favourable predicted drug-likeness characteristics, including high drug-likeness characteristics like high predicted gastrointestinal absorption alongside model-dependent toxicity predictions. Collectively, the results identify (+)-Applanatumol D as a promising multi-target computational candidate for further investigation against CXCR3, CXCR4, and CXCR7. The predicted interactions are consistent with potential receptor binding; however, whether (+)-Applanatumol D acts as an agonist or antagonist and its effects on receptor expression and oral cancer phenotypes require experimental validation.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:4b429edb2b2f97b402383f7c280077121240c0fd","kind":"journals","source":"Journal of applied toxicology","title":"Integrated In Vitro and In Silico Assessment of Dexamethasone-Induced Cytotoxicity and Genotoxicity","url":"https://doi.org/10.1002/jat.70505","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1002%2Fjat.70505","date":"2026-11-16T00:00:00Z","timestamp":1794787200,"categories":["Docking & Screening"],"topic_ids":["docking"],"keywords":["molecular docking","Caco 2"],"matched_keywords":["molecular docking","Caco-2","molecular","protein"],"matched_tags":["docking","permeability","generic_chem","biology"],"doi":"10.1002/jat.70505","external_id":"4b429edb2b2f97b402383f7c280077121240c0fd","pdf_url":null,"code_url":null,"code_host":null,"authors":["Shimaa E. Rashad","Abdel-Hamid A. Haggran"],"journal":"Journal of applied toxicology : JAT","publisher":null,"impact_factor":null,"abstract":"Dexamethasone is a widely used glucocorticoid with potent anti-inflammatory and immunosuppressive properties. Despite its broad clinical application, its cellular toxicity responses under experimental exposure conditions remain incompletely characterized. This study evaluated the cellular responses associated with dexamethasone exposure using complementary in vitro and in silico approaches. Cytotoxicity was assessed in MCF-7, Caco-2, Vero, and HFB4 cell lines using the MTT assay. MCF-7 cells were subsequently selected for apoptosis and cell cycle analysis by flow cytometry based on their greater sensitivity to dexamethasone. DNA damage-associated responses were independently evaluated using the Saccharomyces cerevisiae yeast comet assay in three knockout strains. Complementary computational analyses included UALCAN gene expression profiling, STRING and GeneMANIA protein-protein interaction analyses, and molecular docking of dexamethasone with CASP3 and BCL-2. Dexamethasone produced concentration-dependent reductions in cell viability across the tested cell lines and was associated with increased apoptotic and necrotic populations and altered cell cycle distribution in MCF-7 cells. The yeast comet assay showed increased DNA damage-associated parameters following dexamethasone exposure, with responses varying among the tested knockout strains. Computational analyses provided contextual information regarding apoptosis-associated genes and protein networks, while molecular docking identified predicted structural compatibility between dexamethasone and CASP3/BCL-2. These computational findings are hypothesis generating and do not establish treatment-specific gene regulation, direct target engagement, or causal molecular mechanisms. Overall, dexamethasone elicited multiple cellular responses under the present experimental conditions, including reduced metabolic activity, apoptosis-associated changes, altered cell cycle distribution, and DNA damage-associated responses in an independent yeast model. Further studies using mammalian genotoxicity assays and direct molecular validation are required to establish the underlying mechanisms.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:cba1407a26e2687176840540c42714438f10b528","kind":"journals","source":"Drug and chemical toxicology","title":"Integrative computational and in-vitro evidence implicates BCL2-associated apoptotic signaling in acetyl tributyl citrate-induced microglial toxicity","url":"https://doi.org/10.1080/01480545.2026.2753476","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1080%2F01480545.2026.2753476","date":"2026-11-16T00:00:00Z","timestamp":1794787200,"categories":["Docking & Screening"],"topic_ids":["docking"],"keywords":["Molecular docking","XGBoost","molecular dynamics","IC50"],"matched_keywords":["Molecular docking","XGBoost","Molecular","molecular dynamics","IC50"],"matched_tags":["docking","ensemble","generic_chem","molecular_dynamics","bioactivity"],"doi":"10.1080/01480545.2026.2753476","external_id":"cba1407a26e2687176840540c42714438f10b528","pdf_url":null,"code_url":null,"code_host":null,"authors":["Song-Song Zou","Lei Li","Wen-Jun Xiang","Yang-Lei Li","Fen Wang","Peng-Yu Wei","Yong-Su Lai","Shan-Shan Zou","Li-Li Chen"],"journal":"Drug and chemical toxicology","publisher":null,"impact_factor":null,"abstract":"Acetyl tributyl citrate (ATBC), a synthetic plasticizer, exhibits neurotoxic potential. Microglia sustain brain homeostasis, and their apoptosis and dysfunction are key drivers of Alzheimer's disease (AD) pathogenesis. However, the mechanism connecting ATBC exposure to microglial impairment and AD susceptibility remains unclear. Here, we applied an integrative computational and in-vitro framework to investigate BCL2-centered apoptotic signaling mediating ATBC-induced microglial toxicity. We retrieved ATBC targets from five prediction databases and AD-related genes from GeneCards and OMIM, obtaining 380 overlapping genes. Five hub genes (SRC, TNF, EGFR, BCL2, and ESR1) were screened via four centrality algorithms. Functional enrichment highlighted neurodegenerative disease pathways and microglia-related apoptotic cascades. We built 127 machine-learning models using two AD GEO cohorts (GSE122603, GSE97760). The robust Lasso + XGBoost ensemble (training AUC = 0.994) identified BCL2, SRC, ESR1, and EGFR as predictive features, with SHAP analysis pinpointing BCL2 as the top contributor. Molecular docking predicted favorable ATBC-BCL2 binding (-5.363 kcal/mol), and 100 ns molecular dynamics simulations supported stable complex formation. In BV2 microglia, ATBC dose-dependently reduced cell viability (IC50 = 59.59 μM), downregulated BCL2 expression, and activated caspase-dependent apoptosis. Collectively, our integrated evidence indicates that BCL2-associated apoptotic signaling mediates ATBC-provoked microglial toxicity, suggesting a plausible mechanistic link between ATBC exposure and AD-related signatures. This study proposes a microglia-oriented mechanistic hypothesis for AD susceptibility and offers a multi-omics paradigm for plasticizer risk assessment.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:c9bbe59b5789d3f93849af161a28752fd5a9ef00","kind":"journals","source":"Biomedicine & pharmacotherapy = Biomedecine & pharmacotherapie","title":"Juglone exerts acaricidal activity against Sarcoptes scabiei via dual inhibition of mite mitochondrial complexes III and IV for scabies control","url":"https://doi.org/10.1016/j.biopha.2026.120046","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1016%2Fj.biopha.2026.120046","date":"2026-11-16T00:00:00Z","timestamp":1794787200,"categories":[],"topic_ids":[],"keywords":["molecular docking","enzyme"],"matched_keywords":["molecular docking","molecular","enzyme"],"matched_tags":["docking","generic_chem","biology"],"doi":"10.1016/j.biopha.2026.120046","external_id":"c9bbe59b5789d3f93849af161a28752fd5a9ef00","pdf_url":null,"code_url":null,"code_host":null,"authors":["Li-Xia Dai","Yu-Dong Ma","Hao-Wen Wu","Zi-Le Gong","Qian Liu","Jing Wang","Hong-Wei Hu","Xiao-Lou Miao","Xiao-Rong Yang","Xiao-Fei Shang"],"journal":"Biomedicine & pharmacotherapy = Biomedecine & pharmacotherapie","publisher":null,"impact_factor":null,"abstract":"Scabies is a globally prevalent neglected tropical disease affecting approximately 500 million people. While juglone exhibits promising acaricidal properties, its molecular targets and underlying mechanisms remain largely unknown. This study aimed to elucidate juglone's binding interactions, acaricidal mechanisms, and species specificity using surface plasmon resonance (SPR), RNA-Seq analysis, molecular docking, Western blotting, and enzyme activity assays. The results reveal that juglone demonstrates significant acaricidal activity against Sarcoptes scabiei and Psoroptes cuniculi. Mechanistically, it dual-targets cytochrome b (Cyt b) and cytochrome c oxidase subunit 1 (COX-1) to inhibit mitochondrial respiratory complexes III and IV. Through hydrogen bonding and π-alkyl interactions, juglone induces electron transport collapse, with 40 μg/mL yielding inhibition rates of 48.94% and 66.51% for complexes III and IV, respectively. This respiratory chain dysfunction triggers excessive reactive oxygen species (ROS) production, impairs energy metabolism, and disrupts mitochondrial homeostasis via excessive fission and mitophagy, ultimately causing irreversible structural damage. Furthermore, juglone exhibited distinct species specificity; it significantly inhibited mitochondrial membrane potential and complex activity in C6/36 cells but showed no activity in RAW264.7 cells, indicating that insect cells are particularly sensitive to juglone (i.e., the C6/36 cell line). In conclusion, juglone exerts robust acaricidal effects by disrupting mitochondrial quality control and energy metabolism via complexes III and IV, laying the foundation for its development as a novel alternative acaricidal agent.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:10.1038/s41467-026-79078-4","kind":"journals","source":"Nature Communications","title":"Knowledge-guided machine learning deciphers spatiotemporal dynamics in chemical reactor systems","url":"https://doi.org/10.1038/s41467-026-79078-4","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1038%2Fs41467-026-79078-4","date":"2026-11-16T00:00:00+00:00","timestamp":1794787200,"categories":[],"topic_ids":[],"keywords":[],"matched_keywords":["machine learning","chemical"],"matched_tags":["machine_learning","generic_chem"],"doi":"10.1038/s41467-026-79078-4","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Shiyi Tang","Zhaoyang Duan","Zhou Tian","Wenli Du","Feng Qian"],"journal":"Nature Communications","publisher":"Nature","impact_factor":15.7,"abstract":"","source_metadata":{"short_name":"Nat. Commun.","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"preprints:10.26434/chemrxiv.15012306/v1","kind":"preprints","source":"ChemRxiv","title":"LLaMalysis: Language-guided inverse design of catalyst surfaces with targeted hydrogen adsorption energies","url":"https://doi.org/10.26434/chemrxiv.15012306/v1","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.26434%2Fchemrxiv.15012306%2Fv1","date":"2026-11-16T00:00:00Z","timestamp":1794787200,"categories":[],"topic_ids":[],"keywords":["generative models","inverse design","density functional theory"],"matched_keywords":["generative models","inverse design","density functional theory"],"matched_tags":["generative_model","inverse_design","dft"],"doi":"10.26434/chemrxiv.15012306/v1","external_id":"10.26434/chemrxiv.15012306/v1","pdf_url":null,"code_url":null,"code_host":null,"authors":["Daehyun Kim","Kwangnam Kim","Tuan Anh Pham","Haesun Park"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"Computational catalyst discovery is constrained by vast compositional and structural spaces, whereas atomistic generative models typically rely on engineered conditioning variables. Here we formulate catalyst inverse design as an instruction-following task in which compositional and catalytic requirements are expressed in natural language. We fine-tune an instruction-tuned language model on serialized catalyst slabs from the Open Catalyst 2020 dataset and obtain a generator with validity and uniqueness comparable to catalyst-specific baselines. Reward-weighted regression using machine-learning interatomic-potential feedback increases the number of candidates satisfying the validity criteria and lying within 0.10 eV of the target from 8 to a maximum of 65 per 500 generations. The refined model retains metal-specific controllability for ten metals, and density functional theory calculations of 101 prioritized candidates identify multiple alloy families with promising hydrogen adsorption energies. These results establish natural language as a practical conditioning interface for atomistic catalyst design.","source_metadata":{"date_field":"posted","keywords":[]}},{"id":"journals:10.1021/acs.jcim.6c03293","kind":"journals","source":"Journal of Chemical Information and Modeling","title":"Martini 3 Coarse-Grained\nModels for Steroid Hormones","url":"https://doi.org/10.1021/acs.jcim.6c03293","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1021%2Facs.jcim.6c03293","date":"2026-11-16T00:00:00+00:00","timestamp":1794787200,"categories":["Docking & Screening"],"topic_ids":["docking"],"keywords":[],"matched_keywords":[],"matched_tags":["molecular_dynamics","biology","generic_chem"],"doi":"10.1021/acs.jcim.6c03293","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Tâmela Zamboni Madaloz","Karin dos Santos","Fatmegyul Mustan","Fernando Neiva Nunes","Guilherme Razzera","Manuel N. Melo"],"journal":"Journal of Chemical Information and Modeling","publisher":"ACS","impact_factor":5.3,"abstract":"Steroid hormones are central to the development, metabolism, homeostasis, and reproduction of animals and have been the subject of many evolutionary, ecotoxicological, and human health studies. Coarse-grained molecular dynamics (CG MD) simulations are a powerful tool to study the structural and dynamic behavior of steroid hormones and their interactions. Here, we expand the capabilities of the widely used Martini 3 CG framework by parameterizing and validating a set of 20 steroid hormone models─including both endogenous and synthetic molecules─to enable their study across diverse targets, systems, and fields of interest. Hormone parameterization followed Martini 3 guidelines, while leveraging the similarities with the parent cholesterol molecule, for which a Martini 3 model is available. Notably, we were able to base our parameterization on the hundreds of experimental structures in the protein data bank that contain steroid hormones, instead of relying on atomistic simulations as is typical with Martini. The parameterized models match to within 5% of target molecular volumes and recover the overall hydrophobicity trend of the hormones, although higher hydrophobicities tended to be overestimated. In free simulations of protein-hormone systems, the experimentally known binding poses could be recovered for 89% of the cases, indicating a successful parameterization also from a functional standpoint. Since several Martini 3 protein models were tested, we also incidentally show that the flexibility of the OLIVES and Go̅Martini models makes them best suited for this type of protein–ligand recognition. Finally, models were validated in interaction with membranes by comparison to reference atomistic data. We expect that the availability of our fully validated Martini 3 CG steroid hormone models will contribute to new studies, collaborations, and applications across different scientific fields.","source_metadata":{"short_name":"JCIM","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"journals:fb3af533a71af2f309deeea233b971ba34ea37df","kind":"journals","source":"Frontiers in Pharmacology","title":"Measuring the bile acid receptor FXR: bridging preclinical detection and screening technologies to clinical biomarkers and translational applications","url":"https://doi.org/10.3389/fphar.2026.2011401","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.3389%2Ffphar.2026.2011401","date":"2026-11-16T00:00:00Z","timestamp":1794787200,"categories":["ADMET & Safety"],"topic_ids":["admet"],"keywords":["virtual screening","receptor"],"matched_keywords":["virtual screening","ligand","ligands","molecular","receptor"],"matched_tags":["virtual_screening","generic_chem","biology"],"doi":"10.3389/fphar.2026.2011401","external_id":"fb3af533a71af2f309deeea233b971ba34ea37df","pdf_url":null,"code_url":null,"code_host":null,"authors":["Lei Liu","Bing-Bing Zhang","Kai Sang","Ji-Yong Lin","Ming-Shun Li","Pei-Lin Zhong","Jian-Ping Lai","Liu-Bin Xu","Li Chen","Yu-Xiang Liu","Sheng Chen"],"journal":"Frontiers in Pharmacology","publisher":null,"impact_factor":null,"abstract":"The farnesoid X receptor (FXR; NR1H4 ) is a ligand-activated nuclear receptor that coordinates bile acid, lipid, and glucose homeostasis and has emerged as a validated therapeutic target in cholestatic liver disease and a target in late-stage clinical development in metabolic dysfunction-associated steatohepatitis, though no agent in this class has yet demonstrated a confirmed clinical benefit in that disease. However, the field remains anchored in preclinical and mechanistic research: translation of FXR biology into clinical applications remains limited, and although better measurement is a necessary but not sufficient contribution to that translation—alongside efficacy, safety, and tissue-dependent pharmacology—what we can measure are FXR’s ligand binding, its transcriptional activity, and the downstream outputs it controls. Here we review technologies for the detection and screening of FXR and its ligands, organized along the receptor’s ligand-dependent activation cascade, and trace how these preclinical tools connect to clinical biomarkers that are beginning to guide FXR-targeted therapy. We first describe the molecular basis of FXR detection, then survey cell-free ligand-binding assays, receptor conformational and coactivator-recruitment assays, transcriptional readouts, and structural methods, followed by high-throughput and virtual screening approaches. We then examine the translational gap that separates in vitro measurement from patient-level readouts, and review the clinical biomarkers of FXR activity—led by circulating fibroblast growth factor 19 (FGF19) and the bile acid synthesis intermediate 7α-hydroxy-4-cholesten-3-one (C4)—together with their emerging roles in pharmacodynamic monitoring, patient stratification, and companion diagnostics and imaging. We conclude by scoping the remaining challenges and identifying the measurement questions that would have to be answered before FXR pathway assessment could support individualized therapy—a step that remains aspirational rather than demonstrated.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:30d685f29dbbe90baf91076c8da44e30b122e6b6","kind":"journals","source":"Microorganisms","title":"Mining, Systematic Characterization, and Optimization of Key Elements for Marine-Derived GDP-L-Fucose Biosynthesis","url":"https://doi.org/10.3390/microorganisms14112657","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.3390%2Fmicroorganisms14112657","date":"2026-11-16T00:00:00Z","timestamp":1794787200,"categories":["Docking & Screening","Targets & Structures"],"topic_ids":["docking","targets"],"keywords":["Molecular Docking","enzyme"],"matched_keywords":["Molecular Docking","molecular","enzyme"],"matched_tags":["docking","generic_chem","biology"],"doi":"10.3390/microorganisms14112657","external_id":"30d685f29dbbe90baf91076c8da44e30b122e6b6","pdf_url":null,"code_url":null,"code_host":null,"authors":["Ming-Chun Lin","Long-Yu Fang","Yun-Hao Li","Jun-Feng Chen","Yan-Ru Si","Jia-Hao Zhang","Xin-Ru Xie","Yi-Ting Ju","Lu-Jiang Hao"],"journal":"Microorganisms","publisher":null,"impact_factor":null,"abstract":"GDP-L-fucose serves as a pivotal precursor for the synthesis of fucosylated human milk oligosaccharides (HMOs), exhibiting immense potential in infant nutrition, the development of anti-inflammatory pharmaceuticals, and clinical diagnostics. This study aimed to identify high-performance enzymatic components from marine sources to construct a green and efficient biosynthetic system. Based on thermodynamic analysis (ΔG′° ≈ −72.0 kJ/mol), seven core synthetic enzymes were identified from three marine bacterial strains. Bioinformatics analysis and molecular modeling (AlphaFold/Molecular Docking) suggested that these enzymes possess favorable thermal stability and substrate-binding affinities, with catalytic mechanisms highly conserved relative to their isoenzymes. A series of seven engineered strains were constructed via heterologous expression, followed by the establishment of an in vitro multi-enzyme cascade synthesis system. Experimental results demonstrated that, following optimization of the enzyme ratio, the system achieved a titer of 0.37 g/L GDP-L-fucose, corresponding to approximate molar yields of 11% based on glucose and 32% based on GTP. To address the bottleneck of the cascade system, semi-rational design was employed to engineer the key enzyme Gmd, yielding the superior E208R variant. This variant exhibited a 37% increase in relative enzyme activity while maintaining thermal stability comparable to that of the wild-type enzyme. Consequently, the GDP-L-fucose titer of the in vitro cascade system increased to 0.44 g/L, corresponding to approximate molar yields of 13.4% based on glucose and 37.3% based on GTP. This study not only expands the repertoire of marine-derived GDP-L-fucose synthetic components but also provides a robust technical framework for the large-scale biological manufacturing of high-value functional oligosaccharides.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:10.1021/acsomega.6c09697","kind":"journals","source":"ACS Omega","title":"Mixture-of-Prompts\nfor Parameter-Efficient ADMET Prediction","url":"https://doi.org/10.1021/acsomega.6c09697","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1021%2Facsomega.6c09697","date":"2026-11-16T00:00:00+00:00","timestamp":1794787200,"categories":["ADMET & Safety"],"topic_ids":["admet"],"keywords":["ADMET Prediction","ADMET"],"matched_keywords":["ADMET Prediction","ADMET","drug","molecular","molecule"],"matched_tags":["admet_prediction","admet","generic_chem"],"doi":"10.1021/acsomega.6c09697","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Jian Huang","Jiayi Yin","Boping Li","Bo Lin","Gang Yu"],"journal":"ACS Omega","publisher":"ACS","impact_factor":null,"abstract":"Accurate prediction of absorption, distribution, metabolism, excretion, and toxicity (ADMET) properties is central to modern drug discovery, yet building separate, fully fine-tuned models for each end point is computationally costly and difficult to scale across many heterogeneous tasks. At the same time, recent molecular foundation models provide rich, transferable representations that invite more parameter-efficient adaptation strategies. In this work, we propose a Mixture-of-Prompts (MoP) framework for parameter-efficient single-task ADMET prediction on top of a frozen molecular encoder. For each end point, we keep the backbone fixed and introduce a small set of learnable prompt “experts” together with a lightweight gating network. Given the encoder representation of a molecule, the gate produces a dense mixture over the experts, whose pooled prompt embedding is concatenated with the frozen molecular embedding and fed into a shallow prediction head. This design allows each end point to express task-specific behavior through a compact prompt expert pool, while updating only a small fraction of the total parameters. We instantiate MoP on the standardized PharmaBench ADMET benchmark, training a separate MoP-adapted model for each of the 11 end points under both random and scaffold splits, and further evaluate it on external OpenADMET/ExpansionRx end points. Experiments show that MoP matches or improves upon strong baselines based on head-only adaptation, LoRA, and prefix tuning on many end points, with robust performance under scaffold-split evaluation and informative behavior under external distribution shift. These results suggest that modeling mixtures of prompt experts on top of frozen molecular foundation models provides a flexible and scalable route toward foundation-model-based ADMET modeling.","source_metadata":{"short_name":"ACS Omega","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"journals:870c57226ae49c491ef28cac27e7792b82030e5d","kind":"journals","source":"The Journal of Physical\nChemistry B","title":"Molecular Insights\ninto EGFR Inhibition by Lapatinib\nand Emerging HER2 Mutant Kinase Inhibitors","url":"https://doi.org/10.1021/acs.jpcb.6c03843","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1021%2Facs.jpcb.6c03843","date":"2026-11-16T00:00:00Z","timestamp":1794787200,"categories":["Docking & Screening","Free Energy & MD"],"topic_ids":["docking","freeenergy"],"keywords":["molecular docking","binding free energy","molecular dynamics","Kinase","receptor"],"matched_keywords":["molecular docking","binding free energy","Molecular","molecular dynamics","Kinase","receptor"],"matched_tags":["docking","thermodynamics","generic_chem","molecular_dynamics","biology"],"doi":"10.1021/acs.jpcb.6c03843","external_id":"870c57226ae49c491ef28cac27e7792b82030e5d","pdf_url":null,"code_url":null,"code_host":null,"authors":["K. G. Reddy","Atanu K. Metya"],"journal":"The Journal of Physical\nChemistry B","publisher":null,"impact_factor":null,"abstract":"The epidermal growth factor receptor (EGFR) plays a central role in the progression of multiple cancers, and resistance to tyrosine kinase inhibitors remains a significant therapeutic challenge. In this study, we employ an integrated computational approach combining molecular docking, all-atom molecular dynamics simulations, and the advanced funnel metadynamics (FM) method to investigate the binding mechanisms of the experimentally supported lapatinib-EGFR system and the potential binding behavior of recently approved HER2 mutant inhibitors zongertinib and sevabertinib toward the EGFR kinase domain. Lapatinib exhibits structural stability of the EGFR kinase domain throughout the simulation. The modeled HER2 mutant kinase inhibitors-EGFR complexes also exhibited stable conformations of the kinase domain. Residue-level interaction and binding mode analysis demonstrate that all inhibitors occupy the canonical binding groove and engage key active-site residues. The relative residence times for lapatinib (1.72 ± 0.14 ns) and zongertinib (1.87 ± 0.16 ns) obtained through τRAMD suggest comparable accelerated dissociation times. In contrast, sevabertinib showed a substantially shorter residence time (0.17 ± 0.01 ns), indicating a faster dissociation process and weaker retention within the kinase binding pocket. The absolute binding free energy of lapatinib (−13.7 ± 0.5 kcal mol–1) obtained from FM is in reasonable agreement with the experimental value. Moreover, the binding free energies of zongertinib (−20.0 ± 0.6 kcal mol–1) and sevabertinib (−9.2 ± 0.8 kcal mol–1) follow trends consistent with their relative residence times. Overall, the combined structural and binding-mode analyses suggest that zongertinib may adopt a favorable binding mode within the modeled EGFR complex. These findings provide a computational hypothesis for the potential interaction of zongertinib with EGFR, which requires further experimental validation.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:500d8b675bd273bb7f8862e43f36bd4844e68018","kind":"journals","source":"Journal of Advances in Medical and Pharmaceutical Sciences","title":"Molecular Modelling and QSAR-Guided Design of Aminoquinoline-Pyrimidine Analogues Targeting Mutant Plasmodium falciparum DHFR-TS","url":"https://doi.org/10.9734/jamps/2026/v28i11911","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.9734%2Fjamps%2F2026%2Fv28i11911","date":"2026-11-16T00:00:00Z","timestamp":1794787200,"categories":["Property Prediction","Docking & Screening","ADMET & Safety","Targets & Structures"],"topic_ids":["property","docking","admet","targets"],"keywords":["ADME prediction","pharmacophore screening","QSAR","ADME","molecular dynamics","IC50","pIC50"],"matched_keywords":["ADME prediction","pharmacophore screening","QSAR","quantitative structure-activity relationship","ADME","Molecular","ligand","compounds","compound","molecular dynamics","IC50","pIC50"],"matched_tags":["admet_prediction","pharmacophore","qsar_qspr","admet","generic_chem","molecular_dynamics","bioactivity"],"doi":"10.9734/jamps/2026/v28i11911","external_id":"500d8b675bd273bb7f8862e43f36bd4844e68018","pdf_url":null,"code_url":null,"code_host":null,"authors":["K. Souleymane","Soro Issouf","Kouakou Kouakou Jean-Louis","Fofana Issouf","Niaré Adama","Kré N. Raymond"],"journal":"Journal of Advances in Medical and Pharmaceutical Sciences","publisher":null,"impact_factor":null,"abstract":"Malaria remains a major public health challenge, with an estimated 282 million cases and 610,000 deaths worldwide in 2024. The bifunctional dihydrofolate reductase-thymidylate synthase of Plasmodium falciparum (PfDHFR-TS) is an established antifolate target whose resistance-associated variants motivate ligand-directed discovery. This study combined molecular mechanics, Poisson-Boltzmann-based energy calculations, quantitative structure-activity relationship (QSAR) modelling, pharmacophore screening, in silico ADME prediction, and molecular dynamics to prioritise aminoquinoline-pyrimidine (AQPy) analogues. Twenty-three published AQPy compounds with whole-parasite antiplasmodial IC50 values were divided into a training set of 19 compounds and an external validation set of four compounds. Docking was based on the quadruple-mutant PfDHFR-TS structure (PDB 1J3K). The free-energy descriptor model showed a training-set coefficient of determination of approximately 0.96 and a four-compound external pIC50 root-mean-square error of approximately 0.23. A pharmacophore-guided design space produced 37 screened records corresponding to 33 distinct displayed compound identifiers. The most favourable modelled complexation-energy descriptors were associated with 15-106, 15-104, and 24-1. Calculated ADME descriptors and 200 ns molecular dynamics trajectories provided complementary information about the proposed binding poses and dynamic behaviour. Together, these results identify structural candidates for biochemical target testing and parasite-growth assays, while distinguishing the molecular-model rankings from experimentally measured inhibitory activity.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:c9940566b8f83d35581f5f1231149fa34d792128","kind":"journals","source":"The Journal of Physical\nChemistry Letters","title":"Molecular Recognition\nof 6PGD by Equisetin Perturbs\nRedox Metabolism in Methicillin-Resistant\n Staphylococcus aureus","url":"https://doi.org/10.1021/acs.jpclett.6c03370","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1021%2Facs.jpclett.6c03370","date":"2026-11-16T00:00:00Z","timestamp":1794787200,"categories":[],"topic_ids":[],"keywords":["molecular docking"],"matched_keywords":["molecular docking","Molecular","molecule"],"matched_tags":["docking","generic_chem"],"doi":"10.1021/acs.jpclett.6c03370","external_id":"c9940566b8f83d35581f5f1231149fa34d792128","pdf_url":null,"code_url":null,"code_host":null,"authors":["Wei-Min Xie","Le Dai","Jia-Hua Zhu","Bing-Ji Ma","Miao Liu","Xiao-Long Yang","Xiao-Le Han"],"journal":"The Journal of Physical\nChemistry Letters","publisher":null,"impact_factor":null,"abstract":"Small-molecule recognition of bacterial metabolic enzymes can perturb bacterial physiology, yet how such interactions translate into antibacterial activity remains poorly understood. Herein, we identify equisetin (EQ), a fungal meroterpenoid, as a potent anti-MRSA natural product with pronounced antibiofilm activity and a low propensity for resistance development. EQ induced membrane dysfunction, ATP depletion, reactive oxygen species (ROS) accumulation, and ultrastructural damage. Transcriptomics and qRT-PCR implicated perturbation of pentose phosphate pathway (PPP)-associated redox metabolism, accompanied by reduced gndA transcript levels and 6-phosphogluconate dehydrogenase (6PGD) abundance. Isothermal titration calorimetry demonstrated direct micromolar EQ-6PGD binding, while molecular docking suggested a plausible binding mode. In an MRSA-infected murine wound model, EQ reduced bacterial burden and accelerated tissue repair. Collectively, these findings suggest that membrane dysfunction and PPP-associated redox metabolic perturbation may jointly contribute to EQ-mediated antibacterial activity, identifying 6PGD-associated PPP redox metabolism as a chemically exploitable vulnerability in MRSA.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:10.1002/cmdc.70556","kind":"journals","source":"ChemMedChem","title":"Multidrug Machine Learning Models for Predicting Human Immunodeficiency Virus‐1 Reverse Transcriptase Inhibitor Resistance","url":"https://doi.org/10.1002/cmdc.70556","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1002%2Fcmdc.70556","date":"2026-11-16T00:00:00+00:00","timestamp":1794787200,"categories":[],"topic_ids":[],"keywords":["ChEMBL","XGBoost","random forest"],"matched_keywords":["ChEMBL","neural networks","Machine Learning","XGBoost","random forest","Drug","molecular","compounds"],"matched_tags":["chem_benchmarks","neural_network","machine_learning","ensemble","generic_chem"],"doi":"10.1002/cmdc.70556","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Sumeyye Yilmaz","Huseyin Tunc","Murat Sari","Serdar Durdağı"],"journal":"ChemMedChem","publisher":"Wiley","impact_factor":3.4,"abstract":"Drug resistance remains a major barrier to HIV/AIDS treatment, underscoring the need for rapid evaluation of resistance mechanisms and candidate inhibitors. We present a multidrug machine‐learning framework using drug‐isolate fold‐change (DIF) data to predict resistance to human immunodeficiency virus type 1 (HIV‐1) reverse transcriptase inhibitors (RTIs). Unlike conventional isolate fold‐change (IF) models, DIF integrates viral mutation profiles with inhibitor‐specific molecular information, enabling simultaneous assessment of multiple RTIs, including unseen compounds. Artificial neural networks (ANN), XGBoost (XGB), LightGBM (LGBM), random forest (RF), and support vector regression (SVR) were evaluated. Across all algorithms, class‐specific DIF models consistently yielded higher correlation coefficients and lower mean squared errors than corresponding IF models. XGB‐DIF performed best overall, achieving a mean Pearson correlation coefficient of 0.914 and a mean squared error of 0.064. The models also captured pairwise resistance trends among FDA‐approved RTIs. External validation against an independent ChEMBL dataset yielded an accuracy of 0.824, an AUC of 0.901, and a Pearson correlation coefficient of 0.773 for the best‐performing XGB model. For clinically relevant variants mean absolute errors in logarithmic fold change were as low as 0.057, with near‐perfect resistance ranking for K103N, Y181C, and Y188L. These findings support the robustness and potential utility of the DIF framework.","source_metadata":{"short_name":"ChemMedChem","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"journals:110c56381eeecf837a33527107d10454c020df1d","kind":"journals","source":"Beni-Suef University Journal of Basic and Applied Sciences","title":"Neuropharmacological and hypoglycemic activities of methanolic tubers extract of Pachyrhizus tuberosus: integrated in vivo and in silico evaluation","url":"https://doi.org/10.1186/s43088-026-00857-5","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1186%2Fs43088-026-00857-5","date":"2026-11-16T00:00:00Z","timestamp":1794787200,"categories":["Docking & Screening","ADMET & Safety"],"topic_ids":["docking","admet"],"keywords":["molecular docking","AutoDock Vina","ADME","binding affinity"],"matched_keywords":["molecular docking","AutoDock Vina","ADME","molecular","compounds","binding affinity"],"matched_tags":["docking","docking_software","admet","generic_chem","bioactivity"],"doi":"10.1186/s43088-026-00857-5","external_id":"110c56381eeecf837a33527107d10454c020df1d","pdf_url":null,"code_url":null,"code_host":null,"authors":["M. Arafat","M. Rahman","Proja Barua","Famiha Mirza","Nazia Sultana Niha","Tasfia Tabassum Chowdhury","N. Moni","Mohammed Mubinul Hoque","Arman R. Azad","S. M. Moazzem Hossen"],"journal":"Beni-Suef University Journal of Basic and Applied Sciences","publisher":null,"impact_factor":null,"abstract":"Pachyrhizus tuberosus (Amazonian yam bean) is an edible tuberous legume that has long been used in tropical South America to treat pyrexia, skin diseases, itching, and gastrointestinal and stomach issues. Despite its ethnomedicinal significance, its potential neuropharmacological and hypoglycemic effects have not been well studied. The purpose of this work was to assess the anxiolytic, antidepressant-like, sedative, and hypoglycemic properties of the methanolic extract of P. tuberosus tubers (MTPT) utilizing experimental animal models, supplemented by computer analysis. Pharmacological evaluation was done in Swiss albino mice. Anxiolytic activity was evaluated by Hole-Board Test (HBT) and Elevated Plus Maze (EPM). The antidepressant activity was assessed by the Forced Swimming Test (FST) and Tail Suspension Test (TST). Sedative activities were evaluated by Open Field Test (OFT) and Hole-Cross Test (HCT), and hypoglycemic activity was measured by Oral Glucose Tolerance Test (OGTT). GC–MS analysis was performed for identification of phytochemical constituents of MTPT. The pharmacological potential of the identified secondary metabolites was validated through in silico studies including molecular docking (AutoDock Vina and Discovery Studio 2021), PASS prediction and ADME/T profiling. MTPT showed significant dose-dependent anxiolytic, antidepressant, sedative and hypoglycemic activities. The extract administration at 400 mg/kg (p < 0.001) significantly improved the anxiolytic effects in HBT and EPM, reduced the immobility time in FST and TST (p < 0.001), and greatly inhibited locomotor activity in OFT and HCT, indicating strong sedative effects (p < 0.001). The same dose also significantly decreased blood glucose levels within two hours in the OGTT (p < 0.05). GC-MS/MS identified many bioactive compounds, and molecular docking showed a good binding affinity with docking scores of − 3.1 to − 7.5 kcal/mol, which supports the biological activities observed. The methanol extract of P. tuberosus tubers exhibits promising anxiolytic, antidepressant, sedative, and hypoglycemic effects in experimental models. Both in vivo and in silico findings endorse that P. tuberosus is a potential natural source of bioactive compounds such as (Phthalic acid, Gorlic acid, Benzenepropanoic acid, 2,4-Di-tert-butylphenol3,5-bis(1,1-dimethylethyl)-4-hydroxy-, methyl ester) for developing therapeutic agents against anxiety, depression, insomnia, and hyperglycemia.","source_metadata":{"source":"semantic_scholar"}},{"id":"preprints:10.26434/chemrxiv.15012299/v1","kind":"preprints","source":"ChemRxiv","title":"Novelty-stratified external validation of free in silico ADMET tools","url":"https://doi.org/10.26434/chemrxiv.15012299/v1","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.26434%2Fchemrxiv.15012299%2Fv1","date":"2026-11-16T00:00:00Z","timestamp":1794787200,"categories":["Cheminformatics","Property Prediction","ADMET & Safety"],"topic_ids":["cheminformatics","property","admet"],"keywords":["InChIKey","Random Forest","ADMET","hERG"],"matched_keywords":["InChIKey","Random Forest","ADMET","hERG","compounds","drug","drugs"],"matched_tags":["molecular_notation","ensemble","admet","toxicity","generic_chem"],"doi":"10.26434/chemrxiv.15012299/v1","external_id":"10.26434/chemrxiv.15012299/v1","pdf_url":null,"code_url":null,"code_host":null,"authors":["Sagar Agrawal","Shankar Gejji","Naveen Maurya","Rakesh K. Dixit","Rahul Kumar","Dr. Sagar Agrawal"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"Background and Purpose. Freely available in silico ADMET platforms are used routinely to triage compounds, yet each is validated by its own developers on its own training data, so reported accuracies are neither comparable across tools nor tested on novel chemistry. We asked whether any platform is generally superior, and whether structural novelty affects all endpoints alike. Experimental Approach. SwissADME, pkCSM, ADMETlab 3.0 and ADMET-AI were benchmarked on aqueous solubility, hERG blockade, blood–brain barrier penetration and drug-induced liver injury, using external datasets with identifiable training overlap removed by InChIKey. Compounds were stratified by nearest-neighbour ECFP4 Tanimoto similarity to approved drugs. Performance was computed overall, within strata and along a continuous novelty axis, tested by stratified bootstrap, and compared with a fingerprint Random Forest. Key Results. No platform was best across endpoints; the leader rotated between SwissADME, ADMETlab 3.0 and ADMET-AI. Contrary to expectation, novel chemistry was predicted no less accurately than drug-like chemistry for solubility, hERG or blood–brain barrier penetration (pooled AUROC change −0.04 and −0.02). Only drug-induced liver injury, the one endpoint lacking a structural determinant, lost accuracy (pooled AUROC −0.15), though this endpoint difference was not significant (p = 0.083). The fingerprint baseline matched or beat at least one published platform on every classification endpoint. Conclusion. These data do not support a universal collapse of accuracy on novel chemistry: three of four endpoints show none. What loss there is falls on the endpoint without a structural basis, consistently across tools but not significantly here. Tool selection should be endpoint-specific, and in silico hepatotoxicity predictions confirmed experimentally. © 2026 by the authors. This preprint is distributed under the terms of the Creative Commons Attribution licence (http://creativecommons.org/licenses/by/4.0/). It has not been peer reviewed and is not under consideration at the journal for which it was originally formatted. Keywords","source_metadata":{"date_field":"posted","keywords":[]}},{"id":"journals:1dc0fc28c7f8d3b481c484dad0b5f6ee4b9078fc","kind":"journals","source":"Integrated environmental assessment and management","title":"Persistence revisited: Partitioning rather than media-specific reactivity dominates persistence in the environment","url":"https://doi.org/10.1093/inteam/vjag179","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1093%2Finteam%2Fvjag179","date":"2026-11-16T00:00:00Z","timestamp":1794787200,"categories":[],"topic_ids":[],"keywords":["regression model","chemicals"],"matched_keywords":["regression model","chemicals","compounds","chemical","compound"],"matched_tags":["ml_task","generic_chem"],"doi":"10.1093/inteam/vjag179","external_id":"1dc0fc28c7f8d3b481c484dad0b5f6ee4b9078fc","pdf_url":null,"code_url":null,"code_host":null,"authors":["D. van de Meent","J. Parsons","J. Hermens","G. Schüürmann","D. de Zwart","R. Ebert","J. Struijs","K. D. den Haan","C. van Gestel","K. van Leeuwen","N. van Straalen","P. de Voogt"],"journal":"Integrated environmental assessment and management","publisher":null,"impact_factor":null,"abstract":"Current regulation identifies environmentally persistent chemicals often using thresholds defined for single-media half-lives. This approach ignores the large and even dominating impact of multimedia partitioning on the (bio)availability of compounds that undergo degradation, and accordingly may underestimate the chemical lifetimes by orders of magnitude. This is demonstrated through comparing chemical half-lives in water with overall persistence (Pov) in terms of environmental lifetimes calculated through a steady-state multimedia mass-balance model. While Pov depends on the entrance compartment, Stroebe's τE, the temporal-remote-state expected environmental lifetime, is less sensitive to the initial entry compartment. Thus, τE appears most suited for the prediction of chemical persistence. In this study, a stepwise Ordinary Least Squares (OLS) regression analysis showed that only four chemical properties, log values of Kow (hydrophobicity), kdeg,w and kdeg,a (degradation rate constants in water and air, respectively), and Pv (vapour pressure), are sufficient to estimate τE. But this analysis also showed non-linear error patterns despite overall good statistics, unraveling peculiarities of the non-linear log-log dependence of τE on compound properties. A machine-learning k nearest neighbors (kNN) regression model operating in the property space yields even more accurate and linear results. Application to existing data sets demonstrates the usefulness of this approach. In particular, it shows that the reduction of bioavailability due to multimedia partitioning may cause compounds classified as \"readily biodegradable\" to be identified as persistent when they have sufficiently high hydrophobicity (e.g. phenanthrene with log Kow = 4.5: lifetime > 30 years). The developed tools (SimpleRisk4REACH enabling τE calculation through Stroebe's mathematical procedure; kNN regression model for τE prediction) are made available to the public.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:38fb987962b2eaaac0a180948ce593d7998d70cc","kind":"journals","source":"Plants","title":"Phytochemical Profiling and Preliminary In Vitro Assessment of Thai Traditional Medicine (Yahom Navakot) Extracts on HepG2 Cells, Integrated with ADMET and Molecular Docking Analyses of Selected Identified Compounds","url":"https://doi.org/10.3390/plants15223571","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.3390%2Fplants15223571","date":"2026-11-16T00:00:00Z","timestamp":1794787200,"categories":["Docking & Screening","ADMET & Safety"],"topic_ids":["docking","admet"],"keywords":["Lipinski","Molecular Docking","ADMET"],"matched_keywords":["Lipinski","Molecular Docking","ADMET","Molecular","Compounds","chemical","protein"],"matched_tags":["drug_likeness","docking","admet","generic_chem","biology"],"doi":"10.3390/plants15223571","external_id":"38fb987962b2eaaac0a180948ce593d7998d70cc","pdf_url":null,"code_url":null,"code_host":null,"authors":["Benjamat Bangthong","P. Puthongking","P. Panyatip","Suthida Datham","P. Siriparu","J. Ratha","Muhammad Subhan","C. Yongram","B. Sungthong","T. Padumanonda"],"journal":"Plants","publisher":null,"impact_factor":null,"abstract":"Yahom Navakot (NV) is a traditional Thai polyherbal formulation with a long history of use in traditional medicine. However, its hepatoprotective potential and chemical characteristics remain incompletely characterized. This study investigated the biological activities and chemical profiles of aqueous and methanolic NV extracts using complementary in vitro and in silico approaches. The aqueous and methanolic extracts maintained human hepatocellular carcinoma (HepG2) cell viability above 70% at 25–400 µg/mL. In a hydrogen peroxide (H2O2)-induced oxidative stress model, the methanolic extract at 25 µg/mL, which was selected based on its cytoprotective effect in the H2O2-induced cytotoxicity model, significantly reduced intracellular DCF fluorescence, an indirect indicator of reactive oxygen species (ROS) levels, corresponding to a 37.53% reduction in the H2O2-induced increase in fluorescence (MFI = 8970.00). In lipopolysaccharide (LPS)-stimulated RAW 264.7 murine macrophages, both extracts inhibited nitric oxide (NO) production in a concentration-dependent manner. Gas chromatography–mass spectrometry (GC-MS) tentatively identified 11 compounds meeting the predefined spectral similarity criterion (SI ≥ 90%) across both extracts. These compounds were further assessed using Lipinski’s Rule of Five; in silico absorption, distribution, metabolism, excretion, and toxicity (ADMET) prediction, and molecular docking against Kelch-like ECH-associated protein 1 (Keap1) and inducible nitric oxide synthase (iNOS). γ-sitosterol, β-eudesmol, and macelignan showed relatively favorable predicted binding energies toward both targets. In summary, this study provides an integrated insight into the chemical characteristics of NV extracts alongside their cytoprotective and NO-inhibitory activities.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:1a099f17f04b4573bb0b1bf39c2bb6a3cccfea5e","kind":"journals","source":"Phytotherapy research","title":"Platycodin D Attenuates Lipopolysaccharide-Induced Acute Lung Injury by Modulating CD14-Associated Inflammatory Signaling and Macrophage Phenotypic Responses","url":"https://doi.org/10.1002/ptr.70523","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1002%2Fptr.70523","date":"2026-11-16T00:00:00Z","timestamp":1794787200,"categories":["Docking & Screening"],"topic_ids":["docking"],"keywords":["Molecular docking","pharmacokinetic","receptor"],"matched_keywords":["Molecular docking","pharmacokinetic","Molecular","receptor"],"matched_tags":["docking","admet","generic_chem","biology"],"doi":"10.1002/ptr.70523","external_id":"1a099f17f04b4573bb0b1bf39c2bb6a3cccfea5e","pdf_url":null,"code_url":null,"code_host":null,"authors":["Lian-Ting Liao","Li-Na Liu","Xi Zou","Peng-Fei Shen","Jin-Long Zhao","Jun-Hao Wang","Yu-Fang Shen","Lin-Yu Xiao","Renyikun Yuan","Hong-Wei Gao"],"journal":"Phytotherapy research : PTR","publisher":null,"impact_factor":null,"abstract":"Acute lung injury (ALI) is characterized by excessive pulmonary inflammation and lacks effective targeted pharmacotherapies. This study investigated the protective effects of Platycodin D (PLD) in lipopolysaccharide (LPS)-induced ALI and explored the involvement of CD14-associated inflammatory signaling and macrophage phenotypic regulation. Male C57BL/6 mice were challenged with intratracheal LPS and subsequently treated with PLD by inhalation (5 and 10 mg/kg) or intravenous injection (1 and 2 mg/kg), with dexamethasone (5 mg/kg, i.p.) as the positive control. PLD alleviated histopathological lung injury, reduced the lung index, pulmonary MPO levels, neutrophil infiltration, and pro-inflammatory cytokines, and modulated macrophage-associated inflammatory phenotypes. RNA-seq analysis identified prominent regulation of NF-κB and NOD-like receptor signaling. Molecular docking, cellular co-localization, CETSA, and DARTS analysis provided complementary evidence consistent with a possible interaction involving CD14. PLD suppressed TLR4/MyD88/NF-κB signaling and downstream NLRP3 inflammasome-associated responses in both lung tissue and macrophage models. In macrophages, PLD reduced M1-associated markers and increased M2-associated markers. CD14 knockdown altered these inflammatory and polarization responses, supporting a functional contribution of CD14, although PLD retained regulatory effects on several downstream markers after CD14 silencing. Collectively, these findings indicate that PLD attenuates LPS-induced ALI through modulation of CD14-associated inflammatory signaling and macrophage phenotypic regulation. The efficacy observed following both nebulized and intravenous administration provides a basis for further pharmacokinetic and translational evaluation of PLD.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:10.1021/acs.jctc.6c00213","kind":"journals","source":"Journal of Chemical Theory and Computation","title":"Potential Energy Surfaces in Multilayered Form: Deep\nLearning Based Regression of Analytical Hierarchical Tree Structures\nUsing Unstructured Clouds of Data-Points","url":"https://doi.org/10.1021/acs.jctc.6c00213","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1021%2Facs.jctc.6c00213","date":"2026-11-16T00:00:00+00:00","timestamp":1794787200,"categories":[],"topic_ids":[],"keywords":[],"matched_keywords":["ML","molecular"],"matched_tags":["machine_learning","generic_chem"],"doi":"10.1021/acs.jctc.6c00213","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Z. Qiu","F. Magoulès","D. Peláez"],"journal":"Journal of Chemical Theory and Computation","publisher":"ACS","impact_factor":null,"abstract":"In this work, we introduce analytical hierarchical tree structures, aka multilayered parametric functions, in the context of regression of multidimensional surfaces, such as molecular or interparticle potential energy surfaces. From a technical perspective, our binary trees stem from the ideas underlying the so-called hierarchical Tucker decomposition by replacing columnwise the factors by piecewise polynomials interpolated from randomized multivariate distributions. We refer to our approach as hierarchical Tucker finite basis representation (HT-FBR). Furthermore, our trees can be transformed into a computational graph a recursive representation in terms of operators. Our optimized HT-FBR function is smooth, continuous and differentiable. Using a single-entry rule and our direct index-flow theorem, it can be efficiently evaluated at any system configuration within the definition domain. As a consequence, we show that it is possible to optimize the HT-FBR parameters using clouds of scattered data-points in combination with a deep-learning based approach. Specifically, we show that such optimization process is very efficient when employing a backpropagation-like scheme. To assess the power of our approach, we illustrate these features in the context of multidimensional fitting of potential energy surfaces for two systems: (i) a common benchmark, a (6D) PES for the description of the cis–trans isomerization of HONO and (ii) the 12D coupled diabatic potential energy surfaces of ethene. As reference set for this last case-study, we have employed the database generated through a series of direct dynamics variational multiconfigurational Gaussian trajectories (DD-vMCG) trajectories. As a final remark, our HT-FBR representation is fully compatible with both trajectory- and wavepacket-based methods such as such (ML-)MCTDH. As a consequence, it enables a direct comparison of different nonadiabatical dynamical methods.","source_metadata":{"short_name":"JCTC","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"journals:10.1021/acsomega.6c10178","kind":"journals","source":"ACS Omega","title":"Pre-Training Synthetic\nAccessibility Filtering in\nPyRMD under Transferred Percentile Cutoffs across 95 Target Data Sets","url":"https://doi.org/10.1021/acsomega.6c10178","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1021%2Facsomega.6c10178","date":"2026-11-16T00:00:00+00:00","timestamp":1794787200,"categories":[],"topic_ids":[],"keywords":["virtual screening"],"matched_keywords":["virtual screening","compounds","ligand"],"matched_tags":["virtual_screening","generic_chem"],"doi":"10.1021/acsomega.6c10178","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Fabio Jorge de Nazare Ferreira","Tallita Marques Machado","Fernanda Guilhon-Simplicio"],"journal":"ACS Omega","publisher":"ACS","impact_factor":null,"abstract":"Synthetic accessibility filtering before cross-validation changes both the compounds available for learning and those used to evaluate ligand-based virtual screening models. We assessed SAscore filtering in PyRMD using SA-Filter Impact Benchmarking (SA-FIB), with 95 unfiltered target data sets and 438 filtered conditions across five thresholds. Historical residual-distance percentile cutoffs were transferred, while absolute epsilon values were estimated within each training subset. Mean paired F1 differences were −0.1078, −0.0361, −0.0147, −0.0074, and −0.0011 at SA ≤ 2.5, 3.0, 3.5, 4.0, and 5.0, respectively, relative to the corresponding eligible-cohort baselines. At SA ≤ 5.0, the F1 difference was not significant after Bonferroni correction, while ROC-AUC and PRC-AUC increased by 0.005085 and 0.003060. Combining duplicated processed pools into 92 units retained significance for ROC-AUC but not PRC-AUC. Reoptimizing percentile cutoffs within inner folds for the 92 eligible SA ≤ 3.5 data sets changed mean outer-fold F1 by +0.000294, with 30 improvements, 22 ties, and 40 decreases. Responses varied by target and metric. Historical parameter selection, reused outer folds, and differences between baseline and filtered evaluation populations limit interpretation to sensitivity within this PyRMD protocol. These findings support target-specific validation of synthetic accessibility filters.","source_metadata":{"short_name":"ACS Omega","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"journals:10.1021/acsomega.6c10951","kind":"journals","source":"ACS Omega","title":"Predicting Infinite-Dilution\nActivity Coefficients\nfor Unseen Solutes and Solvents: Accuracy and Uncertainty","url":"https://doi.org/10.1021/acsomega.6c10951","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1021%2Facsomega.6c10951","date":"2026-11-16T00:00:00+00:00","timestamp":1794787200,"categories":[],"topic_ids":[],"keywords":["graph neural network"],"matched_keywords":["neural network","graph neural network","molecular","Chemical"],"matched_tags":["neural_network","gnn","generic_chem"],"doi":"10.1021/acsomega.6c10951","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Jianjun Shi","Shuo Wang","Juan Wang","Haitao Lv"],"journal":"ACS Omega","publisher":"ACS","impact_factor":null,"abstract":"Infinite-dilution activity coefficients (IDACs) provide a thermodynamic basis for comparing solvents, but the reliability of learned predictions depends on the chemistry available during training. We evaluated descriptor models and a Gibbs–Helmholtz graph neural network using 11,832 development observations spanning 541 molecular identities and a reserved literature panel of 2809 binary systems. For ExtraTrees with interaction descriptors (ET-I), mean absolute error in ln γ∞ increased from 0.112 under random splitting to 0.238 for unseen pairs, 0.265 for unseen solutes, and 0.327 for unseen solvents. When both identities were unseen, mean error was 0.673 ± 0.480 across five partitions (median 0.434); only the highest-error partition contained water-solvent test records. The graph model reduced mean error to 0.206 for pair holdout and 0.593 for double holdout, with 70% of its summed double-holdout improvement concentrated in that water-containing partition. Repeated training preserved the lower mean errors relative to ET-I but changed the ordering of models in individual partitions. Sensitivity analyses showed that large replicate disagreements did not account for the water error. Leading physical models were more accurate than ET-I on matched literature systems. Solvent-switch direction agreement increased with predicted contrast magnitude in retrospective comparisons, although candidate ranking uncertainty was not calibrated. Chemical support, partition variability, and contrast-level performance therefore determine how these predictions can inform solvent comparisons beyond familiar systems.","source_metadata":{"short_name":"ACS Omega","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"journals:10.1093/bioinformatics/btag838","kind":"journals","source":"Bioinformatics","title":"ReMoD-Mol: Reliability-Aware Multimodal Learning with Descriptor-Guided Knowledge Distillation for Molecular Property Prediction","url":"https://doi.org/10.1093/bioinformatics/btag838","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1093%2Fbioinformatics%2Fbtag838","date":"2026-11-16T00:00:00+00:00","timestamp":1794787200,"categories":["Cheminformatics","Property Prediction"],"topic_ids":["cheminformatics","property"],"keywords":["SMILES","MoleculeNet","Property Prediction","molecular representation","Molecular Property"],"matched_keywords":["SMILES","MoleculeNet","molecular representation learning","Property Prediction","molecular representation","Molecular","drug","chemical","Molecular Property"],"matched_tags":["molecular_notation","chem_benchmarks","molecular_representation_learning","ml_task","molecular_descriptor","generic_chem","bioactivity"],"doi":"10.1093/bioinformatics/btag838","external_id":null,"pdf_url":null,"code_url":"https://github.com/27167199/ReMoD-Mol","code_host":"GitHub","authors":["Yilin Yang","Shuqi Chen","Qichang Zhao","Cheng Yan"],"journal":"Bioinformatics","publisher":"Oxford","impact_factor":5.4,"abstract":"Motivation Accurate molecular property prediction (MPP) remains a central challenge in drug discovery. Given the limited availability of labeled molecular data, multimodal molecular representation learning offers a promising means of integrating complementary structural information for MPP. Existing multimodal models still face two limitations: (i) common fusion strategies, particularly direct concatenation, can introduce redundant information or signals that are weakly relevant to the target property; and (ii) the effective incorporation of complementary physicochemical priors into multimodal representations remains insufficiently explored. Results Here, we proposed ReMoD-Mol, a reliability-aware multimodal framework that integrated molecular graphs, SMILES sequences and two-dimensional molecular images. Task-adaptive modality calibration learned one set of weights per benchmark from downstream supervision to regulate the relative contributions of the three molecular views. Descriptor- and fingerprint-based teachers further provided task-conditioned chemical priors through train-only, five-fold out-of-fold knowledge distillation. Prototype-guided embedding shaping together with graph-view and SMILES-view consistency objectives regularized the fused representation space. Across nine MoleculeNet benchmark datasets spanning regression and classification, ReMoD-Mol achieved the best mean performance on seven datasets and the second-best on the other two in a scaffold-split comparison with 14 baselines. Representation analyses revealed smoother property-related manifolds and more discriminative class structures, while learned modality coefficients exhibited distinct task-dependent fusion patterns. Teacher-prior and ablation analyses further supported the complementary roles of chemical-prior transfer and multi-view representation regularization. Availability and implementation Source code, processed datasets and implementation details are publicly available at https://github.com/27167199/ReMoD-Mol.","source_metadata":{"short_name":"Bioinformatics","source":"crossref","abstract_source":null,"abstract_quality":null,"code_url":"https://github.com/27167199/ReMoD-Mol","code_status":"found"}},{"id":"journals:a523d4453efec3b6468235f04ab578a98017b64e","kind":"journals","source":"Natural product research","title":"Response surface optimisation of Xiaoyao San polyphenols and anti-inflammatory activity in colonic epithelial cells","url":"https://doi.org/10.1080/14786419.2026.2762084","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1080%2F14786419.2026.2762084","date":"2026-11-16T00:00:00Z","timestamp":1794787200,"categories":["Docking & Screening"],"topic_ids":["docking"],"keywords":["Molecular docking"],"matched_keywords":["Molecular docking","Molecular","chemical"],"matched_tags":["docking","generic_chem"],"doi":"10.1080/14786419.2026.2762084","external_id":"a523d4453efec3b6468235f04ab578a98017b64e","pdf_url":null,"code_url":null,"code_host":null,"authors":["Yu-Xuan Huang","Zhi-Qiang Que","Yuan Cheng","Fang-Ping Zheng","Xue-Qin Chen"],"journal":"Natural product research","publisher":null,"impact_factor":null,"abstract":"Xiaoyao San is a traditional Chinese medicine formula, but its polyphenol extraction and intestinal epithelial activity remain insufficiently defined. We optimised ethanol-reflux extraction by response surface methodology and evaluated the polyphenol-rich fraction in NCM460 cells. The validated optimum was 57% ethanol, a 7.6 mL/g liquid-solid ratio, and 94 min, yielding 7.44 ± 0.15 mg gallic acid equivalents per gram. Xiaoyao San increased cell viability and proliferation, reduced apoptosis and nitrite accumulation, and decreased TNF-α, IL-1β, and IL-6 concentrations. Molecular docking predicted favourable interactions between selected polyphenols and NLRP3, while immunoblotting showed lower NLRP3, caspase-1, and ASC abundance after treatment. The low concentration frequently produced equal or greater effects than the high concentration. These findings connect extraction optimisation with epithelial protection and support chemical standardisation, direct target-engagement studies.","source_metadata":{"source":"semantic_scholar"}},{"id":"preprints:10.26434/chemrxiv.15012291/v1","kind":"preprints","source":"ChemRxiv","title":"Solving Molecular Crystal Structures through Experimental PXRD-Guided Sequential Crystal Structure Prediction","url":"https://doi.org/10.26434/chemrxiv.15012291/v1","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.26434%2Fchemrxiv.15012291%2Fv1","date":"2026-11-16T00:00:00Z","timestamp":1794787200,"categories":["LLMs & Agents"],"topic_ids":["llm"],"keywords":[],"matched_keywords":["neural network","Molecular"],"matched_tags":["neural_network","generic_chem"],"doi":"10.26434/chemrxiv.15012291/v1","external_id":"10.26434/chemrxiv.15012291/v1","pdf_url":null,"code_url":null,"code_host":null,"authors":["Qinyu Qiao","Zikai Xie","Dingrui Fan","Wenjing Cun","Yating Ling","Ruifeng She","Xiaojin Fu","Zhitang Chen","Mingxuan Yuan","Jun Jiang","Chengxi Zhao","Linjiang Chen"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"Structure determination of organic molecular crystals relies heavily on single-crystal X-ray diffraction, yet suitable single crystals are often difficult to obtain. A common alternative uses crystal structure prediction (CSP) to generate candidate structures for subsequent comparison with experimental powder X-ray diffraction (PXRD) patterns. Such trial-and-error searches across multiple space groups incur substantial computational costs. Here, we introduce a structure determination protocol that uses experimental PXRD together with molecular structure as inputs to guide CSP from the outset. A multimodal neural network predicts probabilities for candidate space groups, which are then searched sequentially in descending probability order. Candidate structures are evaluated using PXRD agreement and relative energies, and the search terminates when a candidate satisfies the matching criterion. In a retrospective evaluation of 11 molecular-crystal systems, the protocol returned candidate sets containing structures matching the experimental crystal packing in nine cases. Compared with exhaustive searches across the 22 most common space groups, it reduced the median computational cost of structure determination by 97.9%. We implement this iterative protocol as a fully automated agentic workflow that autonomously coordinates space-group prediction, structure generation, optimization and diffraction matching, enabling crystal structure determination from molecular structure and experimental PXRD.","source_metadata":{"date_field":"posted","keywords":[]}},{"id":"journals:25b5a8e9e3d4f10b42f73817f8fea8e4c6e76839","kind":"journals","source":"Flavour and Fragrance Journal","title":"Study on the Anti‐Fatigue Mechanism of\n Angelica dahurica\n Essential Oil","url":"https://doi.org/10.1002/ffj.70171","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1002%2Fffj.70171","date":"2026-11-16T00:00:00Z","timestamp":1794787200,"categories":["Docking & Screening"],"topic_ids":["docking"],"keywords":["PubChem","molecular docking"],"matched_keywords":["PubChem","molecular docking","molecular"],"matched_tags":["chem_benchmarks","docking","generic_chem"],"doi":"10.1002/ffj.70171","external_id":"25b5a8e9e3d4f10b42f73817f8fea8e4c6e76839","pdf_url":null,"code_url":null,"code_host":null,"authors":["Wei Wang","Zhi-Qing Zhou","Xiao-Long Leng","Shan-Shan Li","Hui-Min Liu","Yu-Rong Sun","Qing-Lei Liu"],"journal":"Flavour and Fragrance Journal","publisher":null,"impact_factor":null,"abstract":"This study aimed to explore the potential anti‐fatigue and neuromodulatory effects of Angelica dahurica essential oil (ADEO). By integrating databases such as GeneCards and PubChem, network pharmacology and molecular docking technologies preliminarily predicted 54 potential core therapeutic targets (including PIK3R1, PIK3CB and PIK3CD) for ADEO against fatigue. In the subsequent weight‐loaded forced swimming test (WL‐FST) mouse model, ADEO inhalation intervention exhibited a trend towards improving exercise endurance and ameliorating fatigue‐related physiological phenotypes. Biochemical and histological evaluations suggest that the essential oil may be involved in regulating energy metabolism (GLU, HG, LA, BUN, LDH), buffering oxidative stress (MDA, SOD), balancing monoamine neurotransmitters (DA, 5‐HT) and stabilizing the hypothalamic–pituitary–adrenal (HPA) axis, while maintaining hippocampal neuronal structure and the expression of neural plasticity markers such as BDNF to a certain extent. To initially evaluate its translational potential in humans, an electroencephalogram (EEG) observation involving 18 volunteers was conducted. However, this section constitutes only preliminary exploratory pre‐experimental data; due to the lack of a solvent blank control group and the small sample size, the suggested potential trend of cortical arousal requires highly cautious interpretation. Overall, the mechanisms and efficacy conclusions deduced in this study remain subject to multiple major limitations: it must be strongly emphasized that the network pharmacology and molecular docking results are purely in silico computational predictions. Due to the complete absence of cell‐level experimental validation targeting PIK3R1, PIK3CD and EGFR, any direct causal link between these theoretical targets and the observed in vivo physiological phenotypes should be interpreted with extreme caution. Furthermore, the human trial has an extremely small sample size and lacks rigorous controls, and systematic long‐term safety evaluations have not yet been conducted. Future comprehensive verifications relying on rigorous molecular biology experiments and large‐scale randomized controlled clinical trials are urgently needed.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:10.1021/acsomega.6c09242","kind":"journals","source":"ACS Omega","title":"Synthesis and Biological\nEvaluation of 1,4-Naphthoquinone-1,2,3-triazole\nHybrids as Potential Antivenom Agents against\n                    Bothrops\n                    spp. Snake Venoms","url":"https://doi.org/10.1021/acsomega.6c09242","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1021%2Facsomega.6c09242","date":"2026-11-16T00:00:00+00:00","timestamp":1794787200,"categories":["ADMET & Safety"],"topic_ids":["admet"],"keywords":["Lipinski","ADME","pharmacokinetic"],"matched_keywords":["Lipinski","ADME","pharmacokinetic","compounds"],"matched_tags":["drug_likeness","admet","generic_chem"],"doi":"10.1021/acsomega.6c09242","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Andresa N. de Lima Silva","Isadora M. G. Andrade","Ruan C. B. Ribeiro","Brenda B. de Oliveira","Mariana de Oliveira Leal Desmarais","Eladio F. Sanchez","Vitor F. Ferreira","Luana da Silva Magalhães Forezi","André L. Fuly","Fernando de Carvalho da Silva"],"journal":"ACS Omega","publisher":"ACS","impact_factor":null,"abstract":"Snakebite envenomation is a neglected tropical disease associated with substantial morbidity and mortality worldwide, highlighting the need for complementary approaches to conventional antivenom therapy. Herein, 12 1,4-naphthoquinone-1,2,3-triazole hybrids (6a–l) were investigated for their ability to inhibit toxic activities induced by the venoms of Bothrops jararaca and B. jararacussu. The compounds were obtained through an ultrasound-assisted copper-catalyzed azide–alkyne cycloaddition (CuAAC) between 2-(azidomethyl)-3-methylnaphthalene-1,4-dione (4) and terminal alkynes (5), affording the derivatives in 54–92% yields. Their effects on venom proteolytic, coagulant, and phospholipase A2 (PLA2) activities were evaluated after incubation with the venoms for 30 min at 37 °C. In addition, in silico ADME and toxicity analyses were performed to assess their physicochemical and pharmacokinetic properties. All compounds complied with Lipinski’s rule of five, although low predicted aqueous solubility was observed for some derivatives and a toxicity alert was identified for 6k. In vitro assays showed that most compounds inhibited venom proteolytic activity by 30–90%, with distinct responses depending on the venom species. Notably, 6i showed no inhibition against B. jararaca venom but reduced proteolytic activity of B. jararacussu by approximately 60%. Compounds 6c, 6g, and 6i also displayed selective inhibition of venom-induced fibrinogen coagulation. Regarding PLA2 activity, 6a completely inhibited the activity of B. jararaca venom, while 6b and 6h produced approximately 92% inhibition; against B. jararacussu, measurable inhibition was observed for 6a, 6e, and 6f. Overall, the results reveal compounds- and venom species-dependent inhibitory profiles and provide biological evidence supporting further investigation of 1,4-naphthoquinone-1,2,3-triazole hybrids as potential scaffolds for complementary approaches to snakebite envenomation.","source_metadata":{"short_name":"ACS Omega","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"journals:c45ee67c5861d5c70be0b3097e9df9848f7bf038","kind":"journals","source":"RSC Advances","title":"Synthesis, crystallographic study, Hirshfeld surface analysis, in vitro antimicrobial evaluation, and computational studies of novel pyrazole–isoxazole hybrid heterocyclic systems","url":"https://doi.org/10.1039/d6ra08878c","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1039%2Fd6ra08878c","date":"2026-11-16T00:00:00Z","timestamp":1794787200,"categories":[],"topic_ids":[],"keywords":["Molecular docking","DFT","ADME"],"matched_keywords":["Molecular docking","DFT","ADME","compounds","Molecular"],"matched_tags":["docking","dft","admet","generic_chem"],"doi":"10.1039/d6ra08878c","external_id":"c45ee67c5861d5c70be0b3097e9df9848f7bf038","pdf_url":null,"code_url":null,"code_host":null,"authors":["Mohammed Chalkha","Aziz Arzine","S. D. Kanmazalp","Khaoula Faiz","Mustapha Igli","E.Bulat Elif Bulat","Y. Sert","B. Louasté","M. Saadi","J. T. Mague","M. Abulmeaty","Pu Xia","M. Aboul-soud","L. El Ammari","M. el Yazidi"],"journal":"RSC Advances","publisher":null,"impact_factor":null,"abstract":"A series of novel pyrazole–isoxazole hybrid heterocycles were synthesized using an efficient method and thoroughly characterized via NMR (1H, 13C) spectroscopy, HRMS and single-crystal X-ray diffraction (XRD). XRD and Hirshfeld surface (HS) analyses provided detailed insights into the solid-state structures, while DFT-based MEP and HOMO–LUMO analyses revealed stable electronic framework with defined interaction sites. The new compounds displayed significant antibacterial activity in vitro against both Gram-positive and Gram-negative bacteria. Among them, 6a showed the highest activity, with a minimum inhibitory concentration (MIC) of 50 µg mL−1 against S. aureus and 80 µg mL−1 against B. subtilis. Moreover, 6a exhibited important antifungal activity against C. albicans, A. niger, A. niger, and F. oxysporum, with MICs ranging from 5 to 8 µg mL−1 and a minimal fungicidal concentration (MFC) of 8 µg mL−1. Molecular docking indicated favorable binding of 6a to DNA gyrase B and lanosterol 14α-demethylase. Despite some ADME limitations, 6a represents a promising scaffold for further antimicrobial optimization.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:23c864bb09db651480789476865dec2f8207ab45","kind":"journals","source":"International Journal of Drug Delivery Technology","title":"Synthesis, Molecular Docking, and Anticancer Activity of Novel Chalcone-Hydrazone Derivatives as Tubulin Polymerization Inhibitors","url":"https://doi.org/10.25258/ijddt.16.80s.93","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.25258%2Fijddt.16.80s.93","date":"2026-11-16T00:00:00Z","timestamp":1794787200,"categories":["Property Prediction","Docking & Screening"],"topic_ids":["property","docking"],"keywords":["Molecular Docking","QSAR","molecular property"],"matched_keywords":["Molecular Docking","QSAR","Molecular","ligand","Compound","compounds","molecular property"],"matched_tags":["docking","qsar_qspr","generic_chem","bioactivity"],"doi":"10.25258/ijddt.16.80s.93","external_id":"23c864bb09db651480789476865dec2f8207ab45","pdf_url":null,"code_url":null,"code_host":null,"authors":["Devakamma Lakumalla","R. Dasgupta","Shahjalal Alam","Junmoni Nath","Syeda Jabeen Fatima","K. Veerabhadrappa"],"journal":"International Journal of Drug Delivery Technology","publisher":null,"impact_factor":null,"abstract":"The colchicine-binding site of tubulin represents an attractive target for the development of microtubuledisrupting anticancer agents. In the present study, a focused series of chalcone-hydrazone derivatives (3a–3j) was designed and investigated using an integrated in silico approach combining molecular property assessment, molecular docking, and ligand-based quantitative structure–activity relationship (QSAR) analysis. The designed derivatives incorporated different para-substituents on the benzylidene ring to examine their influence on tubulin binding and predicted antiproliferative activity. Molecular docking against the colchicine-binding site of human α/β-tubulin (PDB ID: 1SA0) showed favorable binding profiles across the series, with docking scores ranging from −9.1 to −10.8 kcal/mol. Compound 3c, bearing a para-nitro substituent, exhibited the most favorable docking score (−10.8 kcal/mol) and established multiple interactions within the binding pocket, including hydrogenbonding interactions involving Asn258 and Gln247 and additional electrostatic complementarity with Lys254. QSAR analysis further provided favorable activity estimates for the investigated derivatives against MCF-7 and A549 cell lines. Compound 3c showed the lowest estimated IC₅₀ values of 0.8 µM and 1.2 µM, respectively, followed by compound 3j with estimated IC₅₀ values of 1.0 µM and 1.5 µM. Correlation analysis between docking affinity and QSAR-derived pIC₅₀ values showed R² values of 0.808 for MCF-7 and 0.814 for A549 cells. Structure–activity relationship analysis indicated that electron-withdrawing para-substituents generally favored the computational binding and activity profiles of the chalcone-hydrazone scaffold. Overall, compounds 3c and 3j were identified as promising computational leads for further experimental synthesis and biological evaluation as potential tubulin-targeting anticancer agents.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:35d1db756e12aa0729e7adcbaa9a8d56c25cd7d6","kind":"journals","source":"Journal of Solution Chemistry","title":"Targeting EP3 Receptor for Laxative Drug Discovery: Molecular Docking and Virtual Screening of Ricinoleic Acid Analogues","url":"https://doi.org/10.1007/s10953-026-01679-w","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1007%2Fs10953-026-01679-w","date":"2026-11-16T00:00:00Z","timestamp":1794787200,"categories":["Docking & Screening"],"topic_ids":["docking"],"keywords":["Molecular Docking","Virtual Screening","Receptor"],"matched_keywords":["Molecular Docking","Virtual Screening","Drug","Molecular","Receptor"],"matched_tags":["docking","virtual_screening","generic_chem","biology"],"doi":"10.1007/s10953-026-01679-w","external_id":"35d1db756e12aa0729e7adcbaa9a8d56c25cd7d6","pdf_url":null,"code_url":null,"code_host":null,"authors":["S. Gabsi","N. el Houda Ben Fatma","Firas Hfaeidh","Nour Jegham","Mohamed Amine Elkaibi","Meriem Laribi"],"journal":"Journal of Solution Chemistry","publisher":null,"impact_factor":null,"abstract":"","source_metadata":{"source":"semantic_scholar"}},{"id":"preprints:10.26434/chemrxiv.15003660/v2","kind":"preprints","source":"ChemRxiv","title":"The Multi-Objective Computational Design and Prioritization of Novel TSPO PET Radioligands for Improved Inflammation Imaging after Traumatic Brain Injury","url":"https://doi.org/10.26434/chemrxiv.15003660/v2","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.26434%2Fchemrxiv.15003660%2Fv2","date":"2026-11-16T00:00:00Z","timestamp":1794787200,"categories":["Property Prediction","ADMET & Safety","Reaction Informatics"],"topic_ids":["property","admet","reaction"],"keywords":["ASKCOS","retrosynthetic","ADMET","molecular descriptors","molecular dynamics","partition coefficient"],"matched_keywords":["ASKCOS","retrosynthetic","ADMET","molecular descriptors","molecular dynamics","blood-brain barrier permeability","partition coefficient","ligands","molecular","protein"],"matched_tags":["synthesis_planning_tools","retrosynthesis","admet","molecular_descriptor","molecular_dynamics","bbb_prediction","solubility","generic_chem","biology"],"doi":"10.26434/chemrxiv.15003660/v2","external_id":"10.26434/chemrxiv.15003660/v2","pdf_url":null,"code_url":null,"code_host":null,"authors":["Dhruv Subramanian"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"Clinicians employ TSPO PET imaging to quantify the extent of TBI-associated neuroinflammation through radioligands that bind to the central pocket of the human translocator protein. However, many radioligands underperform due to pocket dissociation and low blood-brain barrier permeability. I hypothesize that a targeted cheminformatic pipeline can accelerate the pre-synthetic design of novel TSPO ligands with improved physicochemistry. Eight second-generation ligands were benchmarked using preclinical physicochemical and PET metrics. Emapunil was selected as the modifiable scaffold. Haloalkylation was curated across two functional groups, establishing a 1922-analog dataset. Following prioritization based on molecular descriptors and docking energy, 15 lead candidates (assigned IDs of A#) proceeded to PBPK modeling. Leads were z-scored, and the top three candidates advanced to plasma-input Logan analysis. Consensus ADMET prioritized leads for short-timescale molecular dynamics. The top four analogs were evaluated through a five-replicate 10-ns MD panel along with the reference, and retrosynthetic routes were devised through ASKCOS. A7 showed a significantly improved static and dynamic profile over emapunil. A7 displayed a 5.4-fold increase in partition coefficient and fourfold increase in brain-plasma concentration ratio in the PBPK simulations versus the reference. Logan VT slope estimates were stable for A3, A6, and A7. MD showed A7 retained 28.25±5.94 mean fixed-pocket contacts and ≥10 contacts in 97.96±2.04% of frames. A7’s contact advantage over AC-5216 was statistically significant after Holm correction (p=0.0317). ASKCOS predicted 200 complete synthetic routes for A3, A8, A15, and emapunil with buyable terminals, while A7 displayed an unresolved N-difluoromethylation terminal. Pareto analysis identified A7 as the only nondominated analog. A7 emerged as the strongest radioligand candidate and demonstrates the effectiveness of a cohesive cheminformatic system in PET radioligand development.","source_metadata":{"date_field":"posted","keywords":[]}},{"id":"journals:76c0d858cfe47e989d18385ce34c612cde15355b","kind":"journals","source":"PLOS One","title":"Toddalolactone inhibits the activation of fibroblasts in skin fibrosis by regulating MAOA","url":"https://doi.org/10.1371/journal.pone.0362493","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1371%2Fjournal.pone.0362493","date":"2026-11-16T00:00:00Z","timestamp":1794787200,"categories":["Docking & Screening"],"topic_ids":["docking"],"keywords":["molecular docking"],"matched_keywords":["molecular docking","molecular"],"matched_tags":["docking","generic_chem"],"doi":"10.1371/journal.pone.0362493","external_id":"76c0d858cfe47e989d18385ce34c612cde15355b","pdf_url":null,"code_url":null,"code_host":null,"authors":["Yong-Ping Wang","Shi-Meng Li","Hao Song","Cheng Yang"],"journal":"PLOS One","publisher":null,"impact_factor":null,"abstract":"Skin fibrosis is characterized by excessive deposition of extracellular matrix driven by activated fibroblasts, leading to varying degrees of patient discomfort, deformity, disability, and even death. In vitro studies using mouse primary skin fibroblasts (PSFs) and patient-derived primary keloid fibroblasts (KFs) demonstrated that toddalolactone (Tod) significantly inhibited fibroblast proliferation, migration, and activation and suppressed the TGF-β1/Smad and TGF-β1/non-Smad pathways in these cells. In a bleomycin (BLM)-induced mouse model of skin fibrosis, Tod treatment significantly reduced dermal thickness and collagen deposition in mouse skin, decreased skin hydroxyproline content, and inhibited fibroblast activation as well as the TGF-β1/Smad and TGF-β1/non-Smad pathways in vivo. In a keloid xenograft mouse model, Tod treatment significantly reduced the relative weight of keloids in mice. Histopathological analysis revealed that Tod alleviated the degree of keloid fibrosis and collagen aggregation while inhibiting fibroblast activation in keloid tissue. Furthermore, molecular docking identified monoamine oxidase A (MAOA) as a potential target of Tod. Cellular thermal shift assays and MAOA knockdown using siRNA in PSFs and KFs demonstrated that Tod inhibits the TGF-β1 pathway and suppresses fibroblast activation by inhibiting MAOA. Collectively, the findings reveal that Tod regulates the TGF-β1/Smad and TGF-β1/non-Smad pathways through inhibiting MAOA, thereby affecting skin fibroblast activation and alleviating skin fibrosis.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:bd7f1a29c7fc3b41a098468dc32963246e584c49","kind":"journals","source":"Biophysica","title":"Toxicity and Biomolecular Interactions of Acrylamide with DNA and Serum Albumin: An Integrated In Vitro and In Silico Study","url":"https://doi.org/10.3390/biophysica6060116","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.3390%2Fbiophysica6060116","date":"2026-11-16T00:00:00Z","timestamp":1794787200,"categories":["Docking & Screening"],"topic_ids":["docking"],"keywords":["molecular docking","cytochrome P450","binding affinity"],"matched_keywords":["molecular docking","protein","cytochrome P450","molecular","binding affinity"],"matched_tags":["docking","biology","toxicity","generic_chem","bioactivity"],"doi":"10.3390/biophysica6060116","external_id":"bd7f1a29c7fc3b41a098468dc32963246e584c49","pdf_url":null,"code_url":null,"code_host":null,"authors":["Y. Naji","Maulidiani","Mahboob Alam","N. H. A. Kadir"],"journal":"Biophysica","publisher":null,"impact_factor":null,"abstract":"Acrylamide (AA), a widely used electrophilic vinyl amide in industrial wastewater treatment, has been reported to be genotoxic and carcinogenic in the environment. However, its detailed molecular mechanisms of toxicity and non-covalent biomolecular interactions remain poorly understood. In this work, we studied AA using an integrated in vitro/in silico approach to understand its toxicological profile and interaction behavior. The toxicity evaluation using the Artemia franciscana nauplii lethality test indicates that the nauplii are susceptible to toxicity, with a concentration-dependent decline in survival from 92.86 ± 1.9% at 0.005 μg/mL to 14.29 ± 6.2% at 50 μg/mL after 24 h of exposure. AA toxicity in Artemia is likely due to direct electrophilic action, whereas mammalian toxicity relies largely on cytochrome P450 2E1 (CYP2E1)-mediated conversion to glycidamide. Spectroscopic studies of AA with calf thymus DNA (ctDNA), human serum albumin (HSA), and bovine serum albumin (BSA) showed hyperchromic effects (with a slight bathochromic shift for ctDNA and slight hypsochromic shifts for HSA and BSA) and apparent isosbestic points, indicating direct interaction. The low binding affinity for ctDNA (Kb ~600 M−1) indicates groove-binding or electrostatic association rather than strong intercalation. For HSA and BSA, Stern–Volmer constants (Ksv) were 1.12 × 104 and 1.30 × 104 M−1, respectively, with binding constants (Kb) of 1.22 × 104 and 8.55 × 103 M−1 at 300.15 K. HSA and BSA showed apparent static quenching, whereas the quenching mechanism for the AO–ctDNA system could not be assigned from the quenching rate constant (Kq) alone. Apparent two-temperature thermodynamic indicators suggested favorable, mainly entropy-driven protein binding. The molecular docking study suggested groove binding of AA to ctDNA, with hydrogen bonding (predicted docking score −3.5 kcal/mol), and interactions near the Cys34 region of HSA (−3.7 kcal/mol) and BSA (−4.0 kcal/mol). These weak non-covalent interactions complement, but do not replace, well-established covalent mechanisms of AA toxicity.","source_metadata":{"source":"semantic_scholar"}},{"id":"preprints:10.26434/chemrxiv.15012304/v1","kind":"preprints","source":"ChemRxiv","title":"Transparent Multitask Surface Enhanced Raman Spectral Analysis of Catecholamines with Explainable AI","url":"https://doi.org/10.26434/chemrxiv.15012304/v1","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.26434%2Fchemrxiv.15012304%2Fv1","date":"2026-11-16T00:00:00Z","timestamp":1794787200,"categories":[],"topic_ids":[],"keywords":["convolutional neural"],"matched_keywords":["convolutional neural","machine learning","chemical"],"matched_tags":["neural_network","machine_learning","generic_chem"],"doi":"10.26434/chemrxiv.15012304/v1","external_id":"10.26434/chemrxiv.15012304/v1","pdf_url":null,"code_url":null,"code_host":null,"authors":["Wilson A. Garuba","Aastha Oza","Bhavya Sharma"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"Surface enhanced Raman spectroscopy (SERS) provides a highly sensitive detection of biomolecules; however, interpretation of these high dimensional data is challenging and requires the use of advanced spectra analytical techniques such as machine learning models. These models lack transparency and interpretability, limiting their acceptance in real-world applications. Understanding the rationale behind machine learning predictions would enhance transparency and build confidence among chemical and biomedical researchers. Here, we describe an attention-based multitask convolutional neural network (MCNN) that simultaneously classifies and quantifies three catecholamines (dopamine, epinephrine, norepinephrine) and then apply two explainable AI (XAI) techniques, SHapley Additive exPlanation (SHAP) and Local Interpretable Model Agnostic Explanation (LIME), for interpretation. We demonstrate that a single shared representation supports both identification and quantification of near-identical analytes at 3.6% mean absolute percentage error (MAPE), which is used to estimate the percentage deviation from the true concentration in the regression head. Additionally, the XAI shows the two tasks draw on different spectral regions, with identification based on structure-diagnostic ring modes and quantification linked to intensityscaling and substrate-mediated bands.","source_metadata":{"date_field":"posted","keywords":[]}},{"id":"journals:0dc3cdd606ac59dcaca235b3dabd626f1873a40c","kind":"journals","source":"Antioxidants","title":"Two New Antioxidant Peptides from Ultrasound-Assisted Enzymatic Hydrolysates of Astragalus Seeds: Screening, Identification, Structural Analysis, and Antioxidant Activity Study","url":"https://doi.org/10.3390/antiox15111481","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.3390%2Fantiox15111481","date":"2026-11-16T00:00:00Z","timestamp":1794787200,"categories":[],"topic_ids":[],"keywords":["molecular docking","virtual screening"],"matched_keywords":["molecular docking","virtual screening","molecular","chemical","protein"],"matched_tags":["docking","virtual_screening","generic_chem","biology"],"doi":"10.3390/antiox15111481","external_id":"0dc3cdd606ac59dcaca235b3dabd626f1873a40c","pdf_url":null,"code_url":null,"code_host":null,"authors":["Xue Yan","Ying-Ying Dong","Zhen Xiao","Bing-Mei Wang","Bo Fei","He Teng","Guang-Zhe Li","Shuai Shao","Ming-Ming Yan"],"journal":"Antioxidants","publisher":null,"impact_factor":null,"abstract":"Oxidative stress-mediated cellular damage is involved in various pathological processes. It is of great significance to discover active components with antioxidant protection potential from natural sources. Astragalus [Astragalus membranaceus (Fisch.) Bge. or Astragalus membranaceus (Fisch.) Bge. var. mongholicus (Bge.) Hsiao] is a perennial herbaceous plant in the Fabaceae family, and its anti-tumor, anti-inflammatory, and antioxidant pharmacological activities have been well documented. Notably, Astragalus seeds are rich in protein, but research on the screening and mechanism of antioxidant peptides derived from its seeds remains limited. In this study, Astragalus seed protein hydrolysates (UAMPH) were prepared by ultrasonic-assisted enzymatic hydrolysis. Candidate active peptides were screened using LC-MS/MS identification, multi-database virtual screening, and molecular docking techniques. Two novel pentapeptides, FAVGG and YLGPG, were obtained via solid-phase synthesis. Furthermore, their antioxidant effects were systematically evaluated using in vitro chemical antioxidant assays and an H2O2-induced oxidative damage model in HepG2 cells. The results showed that UAMPH exhibited good structural flexibility, thermal stability, and in vitro gastrointestinal digestibility. Both peptides scavenged various free radicals in a concentration-dependent manner and alleviated intracellular ROS accumulation, restored mitochondrial membrane potential, and inhibited cell apoptosis by suppressing Keap1 expression and increasing the protein expression of Nrf2 and its downstream target HO-1. Concurrently, they increased SOD and CAT activities and GSH levels while decreasing MDA levels and LDH activity. Following intervention with ML385, a selective Nrf2 inhibitor, the peptide-mediated cytoprotective effects were significantly reversed. In conclusion, two previously unreported antioxidant peptides were identified from Astragalus seed protein hydrolysate, and the results suggest that their cytoprotective effects may be associated with activation of the Nrf2-mediated antioxidant response, providing new evidence for the development of traditional Chinese medicinal resources and natural antioxidants.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:10.1021/acs.jmedchem.6c02194","kind":"journals","source":"Journal of Medicinal Chemistry","title":"Ubiquitination-Informed Reinforcement Learning Linker Generator Enables Discovery of Novel FGFR4 PROTAC Degraders","url":"https://doi.org/10.1021/acs.jmedchem.6c02194","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1021%2Facs.jmedchem.6c02194","date":"2026-11-16T00:00:00+00:00","timestamp":1794787200,"categories":[],"topic_ids":[],"keywords":["Reinforcement Learning"],"matched_keywords":["Reinforcement Learning","protein"],"matched_tags":["reinforcement_learning","biology"],"doi":"10.1021/acs.jmedchem.6c02194","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Xunkai Yin","Yihui Cai","Shulan Mei","Songyun Xu","Hanyuan Zhang","Xiaoqian Peng","Wenyu Lu","Zhenzhen Zhu","Wenzhuo Xu","Zhe Zheng","Weijiang Lin","Yang Hu","Lihong Hu","Jian Liu"],"journal":"Journal of Medicinal Chemistry","publisher":"ACS","impact_factor":6.8,"abstract":"Proteolysis-targeting chimeras (PROTACs) represent a promising therapeutic strategy, yet rational linker design is limited by inadequate theoretical guidance. Herein, we report PotentialLinker, a novel reinforcement learning framework for PROTAC linker optimization that uses a normalized weighted harmonic mean (NWHM) score to predict degradation activity. Benchmarking on ternary structures confirms its performance without precomputed protein−protein complexes. In case studies of STAT3/RIPK1, it recapitulated structure−activity relationships (SAR) and recommended appropriate E3 ligases. Then, conducted by PotentialLinker, a potent FGFR4 degrader D11 was obtained (DC50 = 56.23 nM, Dmax = 89.58%). Due to its covalent interacting with the unique Cys552 in FGFR4, D11 possesses excellent FGFR4 selectivity. Furthermore, D11 also demonstrated favorable oral bioavailability (F = 21.3%) and potent antitumor efficacy (TGI = 87.07%) with acceptable tolerability in vivo. Taken together, this framework effectively bridges degradation theory and practical PROTAC design, which can be utilized for the rational design of PROTAC-based therapeutic agents.","source_metadata":{"short_name":"J. Med. Chem.","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"journals:8a467c1d6744511abd8974c658b405ce84702b0f","kind":"journals","source":"Chembiochem","title":"Understanding Nonreacting Substrate Analogs as Potential Scaffolds for Inhibitor Design Against Mycobacterium tuberculosis Cystathionine β-Synthase","url":"https://doi.org/10.1002/cbic.70603","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1002%2Fcbic.70603","date":"2026-11-16T00:00:00Z","timestamp":1794787200,"categories":["Docking & Screening"],"topic_ids":["docking"],"keywords":["Molecular docking","Molecular dynamics","enzyme"],"matched_keywords":["Molecular docking","Molecular","chemical","Molecular dynamics","enzyme"],"matched_tags":["docking","generic_chem","molecular_dynamics","biology"],"doi":"10.1002/cbic.70603","external_id":"8a467c1d6744511abd8974c658b405ce84702b0f","pdf_url":null,"code_url":null,"code_host":null,"authors":["Sainath Polepalli","Anupam Roy","Bapan Mondal","Prasenjit Naskar","Drose Ignatious Shane","Amit Singh","S. Dutta"],"journal":"Chembiochem : a European journal of chemical biology","publisher":null,"impact_factor":null,"abstract":"Mycobacterium tuberculosis cystathionine beta-synthase (MtbCBS), a pyridoxal-5'-phosphate (PLP)-dependent enzyme, regulates sulfur metabolism and redox homeostasis, making it a model enzyme for inhibition studies and to understand scaffold behavior. Amino acids remains as an attractive scaffold for designing novel inhibitors having reduced cytotoxicity and improved properties for cellular uptake. This study integrates computational, biophysical, and structural approaches to provide for a potential scaffold that could be further modified to inhibit MtbCBS. Structure-based sequence alignment of CBS homologs identified seven conserved polar/charged regions critical for PLP-substrate interactions. Molecular docking of polar substrate analogs (threonine, histidine, glutamic acid, and arginine) at the conserved active site revealed distinct binding modes for histidine, glutamic acid, and arginine. Molecular dynamics simulations demonstrated persistent binding of histidine, stabilized by hydrophobic interactions relative to other analogs. Experimental studies revealed histidine's ability to form stable external aldimine adducts, modulating enzyme thermostability (ΔTm), akin to known reacting substrates. Finally, cryo-EM-based structural studies revealed two classes of MtbCBS, one with extra density suggesting the noncovalently bound histidine and the unbound WT MtbCBS maintaining internal aldimine. This suggests that exploration of the chemical modification space of the histidine moiety might provide us with a potential inhibitor against MtbCBS and enzymes of the RTS pathway.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:73658b198397a31c3f856fc85ec1b180d41ced49","kind":"journals","source":"Phytotherapy research","title":"Wogonin Reverses Cisplatin Resistance of Non-Small Cell Lung Cancer by Inducing Ferroptosis and Disrupting Mitochondrial Quality Control: A Potential Role of the HO-1/PPARγ/PGC-1α Axis","url":"https://doi.org/10.1002/ptr.70527","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1002%2Fptr.70527","date":"2026-11-16T00:00:00Z","timestamp":1794787200,"categories":["Docking & Screening"],"topic_ids":["docking"],"keywords":["molecular docking"],"matched_keywords":["molecular docking","molecular"],"matched_tags":["docking","generic_chem"],"doi":"10.1002/ptr.70527","external_id":"73658b198397a31c3f856fc85ec1b180d41ced49","pdf_url":null,"code_url":null,"code_host":null,"authors":["Zi-Cheng Liang","Xia-Yi Fang","Meng-Juan Tao","Xiao-Ning Tan"],"journal":"Phytotherapy research : PTR","publisher":null,"impact_factor":null,"abstract":"Cisplatin resistance remains a major challenge in non-small cell lung cancer (NSCLC) therapy. This study investigated whether wogonin, a natural flavonoid, can reverse cisplatin resistance and explored the underlying mechanisms. The effects of wogonin alone and in combination with cisplatin (DDP) were evaluated in DDP-sensitive (A549) and DDP-resistant (A549/DDP) NSCLC cells using CCK-8, colony formation, wound healing, and Transwell assays. Cell death was characterized by measuring ferroptosis markers (ROS, MDA, GSH, JC-1). Potential targets were identified via RNA sequencing and validated by molecular docking and dynamics simulations. The involvement of the HO-1/PPARγ/PGC-1α axis and mitochondrial quality control was assessed using immunofluorescence, western blotting, qPCR, and electron microscopy, with Zinc Protoporphyrin as an HO-1 inhibitor. In vivo efficacy was confirmed in an NCI-H1975 xenograft model. Wogonin synergized with DDP to inhibit proliferation, migration, and invasion of A549/DDP cells. The combination synergistically induced ferroptosis, characterized by elevated ROS and MDA, depleted GSH, and downregulated SLC7A11. GPX4 was not examined in the current dataset and will be included in follow-up studies. Transcriptomic and in silico analyses identified HO-1, PPARγ, PGC-1α, and SLC7A11 as potential targets. Functionally, the combination suppressed the HO-1/PPARγ/PGC-1α axis and SLC7A11; pharmacological HO-1 inhibition with ZnPP partially reversed these changes. No genetic causality is claimed. Mechanistically, it disrupted mitochondrial quality control by inhibiting biogenesis (reduced mtDNA transcripts, Nrf1/2) and blocking protective mitophagy (altered LC3B localization, reduced Parkin, accumulated p62). In vivo, the combination potently suppressed tumor growth and downregulated the HO-1/PPARγ/PGC-1α axis. These findings suggest that wogonin may reverse cisplatin resistance in NSCLC by inducing ferroptosis and disrupting mitochondrial quality control, potentially through the HO-1/PPARγ/PGC-1α axis. This axis should be considered a working hypothesis pending genetic validation.","source_metadata":{"source":"semantic_scholar"}},{"id":"preprints:2611.21683v1","kind":"preprints","source":"arXiv","title":"When Do Tensor Product Paths Matter? A Group-Theoretic Analysis of Equivariant Interatomic Potentials","url":"https://arxiv.org/abs/2611.21683v1","detail_url":"/chemradar/article?u=https%3A%2F%2Farxiv.org%2Fabs%2F2611.21683v1","date":"2026-11-15T22:20:20Z","timestamp":1794781220,"categories":["ML Potentials"],"topic_ids":["potentials"],"keywords":["NequIP","Equivariant"],"matched_keywords":["NequIP","Equivariant"],"matched_tags":["mlip_tools","equivariance"],"doi":null,"external_id":"2611.21683v1","pdf_url":"https://arxiv.org/pdf/2611.21683v1","code_url":null,"code_host":null,"authors":["Jiawei Wang","Sai Gu","Laurence T. Yang","Junfu Tan","Yemin Yu","Ziheng Lu"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"Tensor-product (TP) layers in equivariant interatomic potentials encode angular interactions through symmetry-allowed paths. Yet local site symmetry can make an allowed path carry no signal after neighbor aggregation. We call this effect Aggregation Cancellation. In a scalar-input first layer, equal-species neighbors on a complete site-symmetry orbit share scalar and radial factors, reducing their angular contribution to a spherical-harmonic orbit sum. If the restricted degree-$l$ representation has no trivial component, this sum and the corresponding path vanish for any learned radial weight. For an ideal $O_h$ octahedral orbit, this removes the $l_\\mathrm{edge}=1,2,3$ classes. Beyond the first layer, the factorisation no longer holds, so cancellation is no longer guaranteed; we instead ask whether affected paths remain less important on average. We test the two levels separately in NequIP: matched-radius TiO$_6$ probes examine first-layer cancellation, whereas a within-model CaTiO$_3$ comparison and six OMat24 models examine the importance bias. We then rank paths on training frames, choose the pruning fraction on validation frames, and evaluate frozen masks on held-out tests. At the highest evaluated fraction (60% of 174 paths), force mean absolute error (MAE) changes are 1.42%, 0.04%, and 1.83% for CaTiO$_3$, OMat24 $O_h$-cubic, and OMat24 low-symmetry data, respectively. The masks imply 67.12-71.94% potential TP-operation reduction and 59.77% potential edge multilayer perceptron (MLP) reduction; these are structural proxies, not measured speedups. The results connect local symmetry to path redundancy and pruning robustness in NequIP; transfer to other architectures remains open.","source_metadata":{"categories":["physics.chem-ph"],"code_status":"absent"}},{"id":"preprints:2611.19812v1","kind":"preprints","source":"arXiv","title":"Reasoning in the loop: antibody lead optimization by probing mechanism and measurement","url":"https://arxiv.org/abs/2611.19812v1","detail_url":"/chemradar/article?u=https%3A%2F%2Farxiv.org%2Fabs%2F2611.19812v1","date":"2026-11-15T05:55:18Z","timestamp":1794722118,"categories":["Design de novo"],"topic_ids":["denovo"],"keywords":["Claude","lead optimization"],"matched_keywords":["Claude","lead optimization","molecules","drug","protein"],"matched_tags":["LLMs","lead_optimization","generic_chem","biology"],"doi":null,"external_id":"2611.19812v1","pdf_url":"https://arxiv.org/pdf/2611.19812v1","code_url":null,"code_host":null,"authors":["Samuel Stanton","Aaron Solomon","Emily Delaney","Melanie Montgomery","Crystal Humphries","Timothy Acker","Ryan Henrici","Charbel Eid","Eddie Abrams","Peyton Greenside"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"Lab-in-the-loop systems, in which a model proposes molecules, automated assays measure their properties and the results retrain the model, are a popular approach to antibody lead optimization. These systems optimize for a specification defined with metrics computed from instrument signals, so their success depends on how well those metrics serve as proxies for the properties a drug candidate needs. When a metric is a poor proxy, neither a better model nor more data will fix it, and identifying the problem usually requires expert review. We report two campaigns run on BigHat Biosciences' Milliner platform, one on an anti-albumin nanobody (VHH) and one on an anti-mesothelin single-chain variable fragment (scFv), in which a research team and a general-purpose AI agent, Claude, brought scientific reasoning into the loop. Over seven rounds and about 1,010 design variants we built libraries from antibody lineage, structure and biophysics, and found from the raw curves that on each scaffold one thermal metric did not measure what its name implied. The libraries were built as controlled comparisons, so the results also show what each mutation contributes to each property. We fit a physical model of the curves, used it to choose metrics suited to each scaffold, and changed the experimental protocol, varying protein concentration and heating rate, to test the model against pre-registered predictions. Both leads improve on their parents in affinity, thermal stability and aggregation while improving or maintaining yield and purity.","source_metadata":{"categories":["cs.LG","cs.AI","q-bio.BM"],"code_status":"absent"}},{"id":"preprints:2611.19793v1","kind":"preprints","source":"arXiv","title":"IANN: A Unified Framework for Equivariant InterAtomic Neural Networks with User-Friendly Implementation","url":"https://arxiv.org/abs/2611.19793v1","detail_url":"/chemradar/article?u=https%3A%2F%2Farxiv.org%2Fabs%2F2611.19793v1","date":"2026-11-15T04:47:47Z","timestamp":1794718067,"categories":["Property Prediction","ML Potentials"],"topic_ids":["property","potentials"],"keywords":["PaiNN","NequIP","MACE","graph neural networks","Equivariant","density functional theory","DFT","PBE"],"matched_keywords":["PaiNN","NequIP","MACE","Neural Networks","Neural Network","graph neural networks","Equivariant","density functional theory","DFT","PBE","molecular"],"matched_tags":["mlip_tools","neural_network","gnn","equivariance","dft","generic_chem"],"doi":null,"external_id":"2611.19793v1","pdf_url":"https://arxiv.org/pdf/2611.19793v1","code_url":"https://github.com/changzhiai/IANN","code_host":"GitHub","authors":["Changzhi Ai","Xueping Qin","Dimosthenis Sokaras","Frank Abild-Pedersen","Johannes Voss"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"Machine-learning interatomic potentials (MLIPs) now approach density functional theory (DFT) accuracy at a fraction of its cost, but the practical benefit is limited by software fragmentation: each state-of-the-art architecture is released as a separate codebase, with its own data format, configuration system, training loop and molecular-dynamics interface. Comparing two architectures on the same dataset, or moving a trained model into a production simulation, therefore requires substantial re-engineering. We introduce IANN (InterAtomic Neural Network), an open-source PyTorch framework that reimplements equivariant graph neural networks - PaiNN, NequIP, MACE, Allegro, EquiformerV2, EquiformerV3 and UMA - behind a unified data object, a unified trainer and a single deployment route, so that the architecture is a one-word change in a configuration dictionary. Forces, stresses and virials are obtained for every architecture from one analytic differentiation of the energy with respect to the edge vectors of the atomic graph, exact under periodic boundary conditions. Distributed training is configured automatically from the SLURM, PBS or MPI environment, so the same script runs unchanged from a workstation to a multi-node GPU allocation. Self-describing checkpoints serve directly as ASE calculators, as ensembles reporting uncertainties, or as TorchScript models run inside LAMMPS through bundled C++ pair styles with single- and multi-GPU inference. Trained on a common dataset of millions of structures, the architectures can be compared under identical conditions. We further release PaiNN foundation models trained on 41.9 million curated DFT structures from MPtrj, Alexandria, OMat24, MatPES, OC20, OC22, OC25, MPr2SCAN and MatPESr2SCAN, spanning the PBE, RPBE and r2SCAN functionals. Additionally, IANN provides an agent interface to facilitate integration with autonomous agents.","source_metadata":{"categories":["cond-mat.mtrl-sci","cs.LG","physics.comp-ph"],"code_status":"found","code_url":"https://github.com/changzhiai/IANN"}},{"id":"journals:10.1038/s41598-026-79484-0","kind":"journals","source":"Scientific Reports","title":"AI-based non-destructive food screening for detecting calcium carbide (CaC2)-treated bananas","url":"https://doi.org/10.1038/s41598-026-79484-0","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1038%2Fs41598-026-79484-0","date":"2026-11-15T00:00:00+00:00","timestamp":1794700800,"categories":[],"topic_ids":[],"keywords":["transformer","chemicals"],"matched_keywords":["transformer","chemicals"],"matched_tags":["transformer","generic_chem"],"doi":"10.1038/s41598-026-79484-0","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Elman Alam","Md Shakhawat Hossain","Md. Tarequl Islam","A. S. M. Ashraf Mahmud","Faisal Tariq"],"journal":"Scientific Reports","publisher":"Nature","impact_factor":null,"abstract":"Artificial fruit ripening using calcium carbide (CaC 2 ) poses a significant food safety concern because prolonged consumption may expose consumers to harmful chemicals. Existing detection methods typically require destructive laboratory analysis, specialized equipment or expert interpretation, limiting their use for rapid on-site screening. Although several AI-based approaches have recently been proposed, they are generally evaluated on individual banana images under controlled conditions, with limited investigation of robustness, computational efficiency and realistic deployment scenarios. This study presents an AI-based framework for the non-destructive detection of calcium carbide-treated bananas using conventional smartphone captured images. Two complementary AI frameworks, Deep Classification with a Frozen Backbone (DCF) and Deep Feature-Based Shallow Classification (DFS), were systematically compared using seven state-of-the-art convolutional and transformer backbones. Unlike previous studies, the proposed framework was evaluated under realistic inspection scenarios containing multiple bananas with different proportions of naturally ripened and calcium carbide-treated fruits, together with comprehensive robustness, explainability and computational efficiency analyses. The DCF-based MobileNet-v2 achieved the highest hold-out classification accuracy of 99.74%, whereas the DFS-based MobileNet-v2​ + Logistic Regression (LR) framework provided the best trade-off between predictive performance and computational efficiency. The MobileNet-v2 + LR framework achieved 95.46% hold-out accuracy, 95.40% 5-fold cross-validation accuracy and correctly classified all 20 images (99.9% mean prediction confidence) in realistic multi-banana inspection, while requiring only 0.65 Giga Floating-Point Operations (GFLOPs) and 0.177 seconds per image. Grad-CAM visualizations further confirmed that the model learned meaningful fruit-specific representations. These findings demonstrate that DFS models combined with smartphone imaging provide an efficient and practical solution for non-destructive food adulteration screening.","source_metadata":{"short_name":"Sci. Rep.","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"journals:ceb15dcebb731ed06cf119512be622a25f1c80b5","kind":"journals","source":"International Journal of Molecular Sciences","title":"Antiviral Activity and Chemical Profiling of Dendronephthya hemprichi Soft Coral Extract Against Human Viruses","url":"https://doi.org/10.3390/ijms272210236","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.3390%2Fijms272210236","date":"2026-11-15T00:00:00Z","timestamp":1794700800,"categories":["Docking & Screening"],"topic_ids":["docking"],"keywords":["molecular docking","DFT","IC50"],"matched_keywords":["molecular docking","DFT","Chemical","compounds","drug","molecular","IC50"],"matched_tags":["docking","dft","generic_chem","bioactivity"],"doi":"10.3390/ijms272210236","external_id":"ceb15dcebb731ed06cf119512be622a25f1c80b5","pdf_url":null,"code_url":null,"code_host":null,"authors":["M. F. Hussein","Emad Manni","Sultan A. Alsahli","Zahrah R. Alrayes","Sondos A. Alhajouj","Y. Elossaily","Rasha M. Alzayed","Mohammed T. A. Salim","A. Fouad","A. Elian","Fahd M. Abdelkarem"],"journal":"International Journal of Molecular Sciences","publisher":null,"impact_factor":null,"abstract":"Marine soft corals represent a promising source of bioactive compounds with potential antiviral applications. In this study, the crude extract of Dendronephthya hemprichi soft coral was evaluated for its antiviral activity against herpes simplex virus type 1 and 2 (HSV-1 and HSV-2), coxsackievirus B4 (CVB4), and human adenovirus 40 (HAdV-40) using in vitro assays. The extract exhibited marked antiviral activity, particularly against HSV-1, with an IC50 of 12.99 ± 0.47 µg/mL (SI = 11.44). In the virucidal protocol, the IC50 was 11.12 ± 0.52 µg/mL, with an SI of 13.35. Relative to the reference drug acyclovir, the extract demonstrated lower potency against both HSV-1 and HSV-2. Gas chromatography–mass spectrometry (GC–MS) analysis revealed a diverse chemical profile dominated by fatty acids, fatty acid methyl esters (FAMEs), terpenoids and sterol derivatives, including oleic acid, 13-Docosenoic acid methyl ester, 9-Octadecenoic acid methyl ester, ethyl isoallocholate, ergosta-5,22-dien-3β-ol acetate, and geranyl isovalerate. Exploratory DFT descriptors and molecular docking of selected metabolites are reported. These analyses are descriptive only and do not explain the measured antiviral activity. The chemical profiling identifies constituents as FAMEs, steroids and terpenoids that may contribute to the observed activity and supports further investigation of D. hemprichi through isolation and mechanistic studies.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:b031c58a43e8e7bc618fb52446de2e2702c9aa21","kind":"journals","source":"International Journal of Biology and Life Sciences","title":"Applications of Deep Learning Algorithms in Protein Structure Prediction and Design","url":"https://doi.org/10.54097/aeebgv76","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.54097%2Faeebgv76","date":"2026-11-15T00:00:00Z","timestamp":1794700800,"categories":["Targets & Structures"],"topic_ids":["targets"],"keywords":["AlphaFold3"],"matched_keywords":["AlphaFold3","Deep Learning","drug","ligands","Protein"],"matched_tags":["ml_protein_ligand","machine_learning","generic_chem","biology"],"doi":"10.54097/aeebgv76","external_id":"b031c58a43e8e7bc618fb52446de2e2702c9aa21","pdf_url":null,"code_url":null,"code_host":null,"authors":["Yi-Jie He"],"journal":"International Journal of Biology and Life Sciences","publisher":null,"impact_factor":null,"abstract":"The three-dimensional configuration of a protein is important for its biological function, which is crucial in biological study and drug development. The progress of deep learning has greatly improved the prediction of protein structures. Especially, the AlphaFold series has increased the prediction precision and extended its use in the simulation of biomolecular complexes and drug discovery. This paper describes the latest progress in protein structure prediction and antibody design based on deep learning. The emphasis is on the enhancement and efficiency of the template-based method and the AlphaFold series. Furthermore, the practical applications of these methods in antibody engineering and drug discovery are illustrated by considering specific examples such as AbEgDiffuser for simultaneous sequence and structure design, and AlphaFold3 for the choice of covalent ligands. Experimental data indicate that deep learning can precisely forecast the protein structures and effectively improve the design and screening of potential substances, thereby offering new possibilities for rational drug design. In the future, more research should concentrate on predicting the conformational changes of proteins, designing novel antibody frameworks from scratch, and obtaining multiple drug-like properties at the same time.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:bb430e9a376fa6af43e0f48231af6bd73ced6019","kind":"journals","source":"Advanced Journal of Scientific Research","title":"Computational Evaluation of Miconazole and Its Proposed Cu(II) and Co(II) Complexes against PBP2a: Molecular Docking and ADME Predictions","url":"https://doi.org/10.67364/hy069055","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.67364%2Fhy069055","date":"2026-11-15T00:00:00Z","timestamp":1794700800,"categories":["Docking & Screening"],"topic_ids":["docking"],"keywords":["TPSA","Molecular Docking","ADME"],"matched_keywords":["TPSA","Molecular Docking","ADME","Molecular","compounds","molecules","protein"],"matched_tags":["drug_likeness","docking","admet","generic_chem","biology"],"doi":"10.67364/hy069055","external_id":"bb430e9a376fa6af43e0f48231af6bd73ced6019","pdf_url":null,"code_url":null,"code_host":null,"authors":["Iman Jaaffer"],"journal":"Advanced Journal of Scientific Research","publisher":null,"impact_factor":null,"abstract":"Methicillin-resistant Staphylococcus aureus (MRSA) remains a major clinical problem because the mecA-encoded penicillin-binding protein 2a (PBP2a) sustains cell-wall synthesis in the presence of β-lactams. Miconazole, an imidazole antifungal, has reported in vitro activity against MRSA, and coordination of azoles to transition metals has been associated with enhanced antimicrobial activity. In this in silico study, free miconazole and its hypothetical Cu(II) and Co(II) complexes were docked against PBP2a (PDB ID: 1VQQ) and the docking was repeated to assess reproducibility. Initial docking scores were −7.2 (free miconazole), −7.4 (Cu complex) and −7.2 kcal/mol (Co complex); repeated docking gave −7.1, −6.9 and −7.0 kcal/mol, respectively, i.e. fluctuations of ≤ 0.5 kcal/mol, with an RMSD of 1.397 Å between runs for the Cu system. SwissADME predicted high gastrointestinal absorption for all three compounds, a decrease in TPSA from 27.05 to 15.71 Å² upon metal coordination, and no P-glycoprotein substrate behaviour. These values are computational predictions only: SwissADME is parameterised for organic molecules, so the descriptors of the metal complexes should be treated as indicative, and the docking scores are not experimental evidence of PBP2a inhibition. The results justify, but do not replace, synthesis and antibacterial testing of the complexes.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:10.1038/s41598-026-78139-4","kind":"journals","source":"Scientific Reports","title":"Computational exploration of Hemidesmus indicus phytochemicals targeting HSP90 in ovarian cancer through molecular docking, structural dynamics and free energy landscape","url":"https://doi.org/10.1038/s41598-026-78139-4","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1038%2Fs41598-026-78139-4","date":"2026-11-15T00:00:00+00:00","timestamp":1794700800,"categories":["Docking & Screening"],"topic_ids":["docking"],"keywords":["molecular docking","free energy"],"matched_keywords":["molecular docking","free energy","molecular"],"matched_tags":["docking","thermodynamics","generic_chem"],"doi":"10.1038/s41598-026-78139-4","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Balasubramanian Sivaprakasam","Manikandan Jayaraman","Soundar Rajan Kulandhaivel","Muthukrishnan Arun","Stalin Antony","Jeyakanthan Jeyaraman","Pandiyan Muthuramalingam","Hyunsuk Shin","Joseph Sahayarayan Jesudass"],"journal":"Scientific Reports","publisher":"Nature","impact_factor":null,"abstract":"","source_metadata":{"short_name":"Sci. Rep.","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"preprints:10.26434/chemrxiv.15012220/v1","kind":"preprints","source":"ChemRxiv","title":"De Novo Design of Blood-Brain Barrier-Penetrant, P-Glycoprotein-Evading EGFR Inhibitors and Hypoxia-Activated Trapping Analogs for Glioblastoma","url":"https://doi.org/10.26434/chemrxiv.15012220/v1","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.26434%2Fchemrxiv.15012220%2Fv1","date":"2026-11-15T00:00:00Z","timestamp":1794700800,"categories":["Cheminformatics","LLMs & Agents"],"topic_ids":["cheminformatics","llm"],"keywords":["TPSA","De Novo Design","SELFIES","XGBoost","LLM","regression model","Receptor","pIC50","IC50"],"matched_keywords":["TPSA","De Novo Design","SELFIES","XGBoost","LLM","regression model","protein","Receptor","pIC50","IC50","drugs","molecular","molecules","molecule","drug","Blood-Brain Barrier penetration"],"matched_tags":["drug_likeness","de_novo_design","molecular_notation","ensemble","LLMs","ml_task","biology","bioactivity","generic_chem","bbb_prediction"],"doi":"10.26434/chemrxiv.15012220/v1","external_id":"10.26434/chemrxiv.15012220/v1","pdf_url":null,"code_url":null,"code_host":null,"authors":["Kathir Murugesan"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"Background: Most Glioblastoma Multiforme (GBM) tumors about 50% grow because of a very overreactive protein called EGFR (Epidermal Growth Factor Receptor). Existing drugs like Erlotinib and Gefitinib work great for lung cancer, but they are not effective for brain cancer and subsequently fail for three main reasons. First, they are too water-attracting (polar, with TPSA > 74 Å2), and second, they are very heavy with a molecular weight above 390 Da. Due to all of these factors, they are unable to cross the Blood-Brain Barrier (BBB). Third, even if they get inside, the brain has security systems like P-glycoprotein (P-gp), which pushes them back outside into the blood, thereby wasting their potential inside the brain. Methods: In order to solve this without the use of an expensive lab, we came up with an automated cloud pipeline on GitHub Actions in collaboration with an LLM (Google Gemini). There are four tools that we used: 1. SELFIES: A system that mutates molecules with proper valency and bonding rules to make them 100% chemically valid. 2. Adaptive Mutation Scaling (σ / sigma): A controller where small tweaks are made when improving a winning molecule, and if the molecule is stuck or not what we are looking for, it makes a bigger jump to get a new mutation. 3. XGBoost: A regressor model trained on real lab data (ChEMBL203) with a similarity guardrail. 4. 3D Protein Docking (PDB: 1M17): A 3D check to test how each molecule actually fits inside the human EGFR crystal structure for Glioblastoma. In total, we completed 1,417,569 generations and the engine tested 4,317,102 unique molecules. Results: There were five reward-hacking measures taken by the regression model and mutator, which were fixed via our iterative human-in-the-loop system: 5. Self-Training Echo Chambers: First, we stopped the AI from feeding its own unverified guesses back into its training data, because those unverified guesses were providing high score results that are practically impossible (pIC50 > 150), which is what we called echo chambers. 6. Unstable Sulfur Triple Bonds: Second, we blocked unstable sulfur triple bonds (ClS#CO...) that would attack healthy cells even though they were yielding a high score in both Blood-Brain Barrier penetration and pIC50 score. 7. Flat Double-Bond Tails Crashing in 3D: Third, we caught flat double-bond tails (=CCCO and NCCC=C1...) that looked good in 2D math but crashed into the roof (Leu694) of the 3D protein pocket, which means they cannot actually attach properly to the Glioblastoma EGFR protein. 8. Tightly Bent Small Rings: Fourth, we blocked tightly bent 3- and 4-membered small rings ([r3] and OC1=NC=C1), because carbon and nitrogen bonds naturally sit at 109.5° to 120°, and forcing them into 60°–90° angles creates severe angle strain (Baeyer strain) that makes them snap open in water. 9. Reactive Chlorine Warheads: Fifth, we filtered out toxic chlorine warheads (-CH2Cl), which are similar to sulfur triple bonds where they could actually harm healthy cells even though they provided good theoretical results. After all of these conversions and fixing these loopholes, we finally got a winner, which is Lead 1 (CCc1c(Nc2cccc(F)c2)ncnc1N1CCCO1). This engine discovered a family of compact, lightweight pyrimidine molecules ranging from 268 to 309 Da that had strong EGFR blocking power (IC50 = 5.20–6.92 nM, pIC50 = 8.16–8.28) and strong 3D binding (ΔG = -7.54 to -8.90 kcal/mol). For further verification, we tested it on the SwissADME server, which confirmed that our Lead 1 can actually cross the Blood-Brain Barrier (TPSA = 50.28 Å2), escapes the P-gp pump (P-gp substrate: No), does not block liver enzymes (CYP3A4 inhibitor: No), and has zero toxicity alerts (0 PAINS / 0 Brenk alerts). This was our result for Lead 1 after the verifications and filters that we passed all the molecules through. How It Reacts in the Body (Our Two New Solutions): When looking at how Lead 1 reacts in the body, we came across two solutions that address the whole problem: 10. The \"Trojan Horse\" Trapping Mechanism (Lead 1 → Lead 1-M): Basically, Lead 1 has a closed 5-membered ring (N1CCCO1) with a weak nitrogen-oxygen (N–O) single bond that hides its polar charges (HBD = 1, TPSA = 50.28 Å2) so it can sneak across the Blood-Brain Barrier. Because its polar charges are masked inside this closed ring, it easily sneaks across the Blood-Brain Barrier and enters an oxygen-starved (hypoxic) brain tumor. Once inside, the reductase enzymes present in the tumor cut the weak N–O bond open, which releases -36.40 kcal/mol of energy (MMFF94s) and turns it into another molecule that we named Lead 1-M (CCc1c(Nc2cccc(F)c2)ncnc1NCCCO — where \"-M\" stands for the active ring-opened Metabolite of Lead 1). Once opened, its polar surface area jumps to 70.07 Å2 and its hydrogen-bond donors jump from 1 to 3. As a result, this active drug gets trapped almost permanently inside the cancer cell while improving its 3D fit inside the EGFR pocket by 39.3%. 11. Backup Molecules for Mutations: In case the tumor mutates its pocket or the liver opens the ring too early due to liver enzymes, we have a backup team from our own database: the","source_metadata":{"date_field":"posted","keywords":[]}},{"id":"journals:bba959bdc4e8694fbc8793232c419ee53405c55a","kind":"journals","source":"Journal of the science of food and agriculture","title":"De novo transcriptome and immunoinformatic analyses identify putative allergen-associated proteins in Oyster mushrooms","url":"https://doi.org/10.1002/jsfa.71233","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1002%2Fjsfa.71233","date":"2026-11-15T00:00:00Z","timestamp":1794700800,"categories":["Docking & Screening"],"topic_ids":["docking"],"keywords":["Molecular docking"],"matched_keywords":["Molecular docking","molecular","Chemical"],"matched_tags":["docking","generic_chem"],"doi":"10.1002/jsfa.71233","external_id":"bba959bdc4e8694fbc8793232c419ee53405c55a","pdf_url":null,"code_url":null,"code_host":null,"authors":["Rashi Chugh","Vikas Handa","A. Upadhyay"],"journal":"Journal of the science of food and agriculture","publisher":null,"impact_factor":null,"abstract":"BACKGROUND Edible mushrooms, including Pleurotus florida and Lentinus sajor-caju, are valued for their nutritional and medicinal properties; however, their allergenic potential remains poorly characterized at the molecular level. The aim of this study was to identify putative allergen-associated proteins using an integrated omics approach combining high-throughput RNA sequencing, de novo transcriptome assembly, differential gene expression (DGE) analysis, and immunoinformatics. RESULTS Transcriptome sequencing and de novo assembly generated transcriptomic resources for both mushroom species. The assembled transcripts were searched against curated-allegen database to shortlist allergen-associated candidates, which were further cross-referenced with the list of DGE genes showing upregulated expressions. Epitope prediction and conservancy analyses were conducted to characterize potential immunogenic regions and assess their conservation across homologous sequences. Molecular docking was used to investigate interactions between selected predicted epitopes or proteins and immunoglobulin E (IgE) antibodies, and in silico trypsin digestion analysis was performed to assess the predicted digestive stability of selected candidates. CONCLUSION This study establishes a transcriptome-based resource of putative allergen-associated proteins in P. florida and L. sajor-caju. The integration of differential gene expression analysis with immunoinformatics and structural analyses provides insights into potential mushroom allergenicity and a foundation for future experimental validation and food safety assessment. © 2026 Society of Chemical Industry.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:2b7e760b82f6a7d94cb67469a5bc09df5a1b74da","kind":"journals","source":"International journal of biological macromolecules","title":"Electrospun potato protein-based composite nanofibers with core-shell structure for curcumin delivery and fruit preservation","url":"https://doi.org/10.1016/j.ijbiomac.2026.155334","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1016%2Fj.ijbiomac.2026.155334","date":"2026-11-15T00:00:00Z","timestamp":1794700800,"categories":["Docking & Screening"],"topic_ids":["docking"],"keywords":["Molecular docking","molecular dynamics"],"matched_keywords":["Molecular docking","molecular","molecular dynamics","protein"],"matched_tags":["docking","generic_chem","molecular_dynamics","biology"],"doi":"10.1016/j.ijbiomac.2026.155334","external_id":"2b7e760b82f6a7d94cb67469a5bc09df5a1b74da","pdf_url":null,"code_url":null,"code_host":null,"authors":["Xing-Zhi Wang","You-Ning Zou","Nian-Hang Chen","Yuan-Ping Ye","Shu-Jin Liao","Ya-Dong Zhen","Zi-Cong Chen","Lu Jiang","Su-Qing Zhao"],"journal":"International journal of biological macromolecules","publisher":null,"impact_factor":null,"abstract":"Low-molecular-weight proteins offer great promise in the field of sustainable food packaging but suffer from poor mechanical and water stability. This study proposes a multi-level synergistic strategy that utilizes composition, cross-linking, and core-shell structure design to transform low-cost potato protein (PP), a byproduct, into high-performance functional nanofibers. By optimizing the PP: Pul mass ratio (2:8) and applying glutaraldehyde (GA) cross-linking treatment, the performance of the PP-based nanofibers was significantly improved, with the best tensile strength of 0.269 MPa, an elongation at break of 6.5%, a higher water contact angle, and enhanced water stability. Molecular docking and molecular dynamics simulations suggested the presence of a dense hydrogen-bond network between PP and Pul, which may contribute to the enhanced mechanical properties of PP/Pul nanofibers. Furthermore, PP/Pul composite nanofibers with a core-shell structure were prepared via coaxial electrospinning to achieve the efficient encapsulation and controlled release of photosensitive curcumin. The curcumin retention rate of core-shell nanofibers (PP-2:8 Cur) reached 46.2% after 24 h of UV irradiation, with release behavior conforming to first-order diffusion and sustained release, with significantly reduced burst release. Strawberry preservation experiments demonstrated that curcumin-loaded core-shell-structured nanofibers could significantly inhibit water loss, color deterioration, and organic acid depletion in strawberries, and effectively extend their shelf life from 3 to 7 days at room temperature. This study provides a paradigm for upgrading low-molecular-weight plant proteins into functional food packaging materials through a multi-level synergistic strategy.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:10.1038/s41598-026-79957-2","kind":"journals","source":"Scientific Reports","title":"Explainable IoT intrusion detection via bio-inspired encoding and interpretable AI models","url":"https://doi.org/10.1038/s41598-026-79957-2","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1038%2Fs41598-026-79957-2","date":"2026-11-15T00:00:00+00:00","timestamp":1794700800,"categories":[],"topic_ids":[],"keywords":["recurrent neural","RNN"],"matched_keywords":["recurrent neural","RNN","Molecular","protein"],"matched_tags":["neural_network","generic_chem","biology"],"doi":"10.1038/s41598-026-79957-2","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Thaer AL Ibaisi","Muhammad Kazim","Mujeeb Ur Rehman","Mustafa Kaiiali","Stefan Kuhn"],"journal":"Scientific Reports","publisher":"Nature","impact_factor":null,"abstract":"We introduce an explainable Internet of Things (IoT) intrusion detection framework built on a bio-inspired Amino Acid encoding of network flows. The flow features are first transformed with the Yeo–Johnson transformation, the most informative ones are selected, and their values are encoded as Amino Acid sequences, which are then summarized by ten protein structural properties (e.g., $$\\beta $$ -Turn, Isoelectric Point and Molecular Weight). In this way every flow is described by the same compact set of ten properties, whatever the feature set of the original dataset. The models are explained with SHAP, LIME and ELI5, complemented by Captum Integrated Gradients and Alibi anchors for the neural models, and the attributions are linked back to the raw traffic features through measured correlations. On a balanced 50,000-flow subset of CICIoT2023 that covers all 33 attack classes, most of the 12 classical and 5 deep models reach 98.4–98.5% accuracy on a held-out test set (MCC up to 0.97). The recurrent neural network (RNN), used for the in-depth explainability study, reaches 98.5% accuracy with a false-alarm rate of 0.02%, has fewer than 15,000 parameters and needs about 0.1 ms per flow on a CPU. An ablation experiment shows that encoding the selected features keeps detection close to the full 39-feature raw representation (MCC 0.970 against 0.979), while giving a fixed ten-property input that can be explained. All explanation methods agree that $$\\beta $$ -Turn and Isoelectric Point drive most decisions, and both properties correlate strongly with the packet count and packet size of the flows.","source_metadata":{"short_name":"Sci. Rep.","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"journals:6331605a13bf19f5809ca26bb9421a6db403f2e3","kind":"journals","source":"IEEE Internet of Things Journal","title":"FedGE-GATNet: Federated Wearable IoMT Intelligence for Explainable Diabetes Severity Prediction Using Biomedical Graph Learning","url":"https://doi.org/10.1109/JIOT.2026.3729647","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1109%2FJIOT.2026.3729647","date":"2026-11-15T00:00:00Z","timestamp":1794700800,"categories":[],"topic_ids":[],"keywords":[],"matched_keywords":["deep learning","drug","protein"],"matched_tags":["machine_learning","generic_chem","biology"],"doi":"10.1109/JIOT.2026.3729647","external_id":"6331605a13bf19f5809ca26bb9421a6db403f2e3","pdf_url":null,"code_url":null,"code_host":null,"authors":["Jing-Jing Tang","Nan Wang"],"journal":"IEEE Internet of Things Journal","publisher":null,"impact_factor":null,"abstract":"Diabetes severity prediction has become increasingly important in smart healthcare systems due to the rapid growth of wearable sensing devices, remote patient monitoring platforms, and distributed clinical data repositories. However, existing prediction frameworks often struggle to jointly model physiological signals, pharmacological interactions, and heterogeneous biomedical relationships while preserving patient privacy across decentralized healthcare environments. This work proposes FedGE-GATNet, a federated graph-enhanced attention framework for distributed diabetes severity prediction using wearable physiological monitoring, clinical biomarkers, and network pharmacology-driven biomedical interaction learning. Unlike conventional deep learning approaches that primarily rely on isolated clinical features, the proposed framework integrates ghost feature generation, graph attention propagation, biomedical interaction-aware representation learning, and adaptive federated optimization to capture patient–drug–protein–pathway dependencies under distributed healthcare settings. In addition, an adaptive Gorilla–Rabbit optimization (AGRO) strategy is introduced for stable federated hyperparameter tuning, while a new biomedical graph reliability index (BGRI) metric is designed to evaluate graph-preserving biomedical prediction consistency. Experiments were conducted on the real-world diabetes network pharmacology dataset (RDNPD) containing 466 605 patient records together with PIMA, MIMIC-III, and OhioT1DM datasets. FedGE-GATNet achieved 97.6% accuracy, 97.4% $F1$ -score, and 0.946 BGRI on RDNPD while consistently outperforming existing federated healthcare and graph-based baseline models across all datasets. The framework further demonstrated robust cross-dataset generalization, stable federated convergence, strong interpretability, and reduced computational complexity. The obtained results confirm the effectiveness of federated biomedical graph intelligence for scalable and explainable diabetes severity prediction in distributed smart healthcare systems.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:ce36e8616134e3f565046a752050918804474819","kind":"journals","source":"Advanced Science","title":"GPCR‐Mediated Regulation of Cytochrome P450 Drives Sulfoxaflor Resistance in Aphis gossypii Glover","url":"https://doi.org/10.1002/advs.78754","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1002%2Fadvs.78754","date":"2026-11-15T00:00:00Z","timestamp":1794700800,"categories":["Docking & Screening"],"topic_ids":["docking"],"keywords":["Molecular docking","Cytochrome P450","GPCR","receptor","enzyme"],"matched_keywords":["Molecular docking","Cytochrome P450","molecular","GPCR","protein","receptor","enzyme"],"matched_tags":["docking","toxicity","generic_chem","biology"],"doi":"10.1002/advs.78754","external_id":"ce36e8616134e3f565046a752050918804474819","pdf_url":null,"code_url":null,"code_host":null,"authors":["Hai-Xiang Lv","Xu-Chao Li","Cheng-Zhi Yang","Zhi-Min Guo","Yang Huang","Shun He","Hu Wan","Jian-Hong Li","C. Bass","Kang-Sheng Ma"],"journal":"Advanced Science","publisher":null,"impact_factor":null,"abstract":"Rapid evolution of insecticide resistance threatens agricultural productivity. While the fundamental molecular and biochemical mechanisms underlying resistance are well documented, the regulatory processes governing resistance‐associated genes, particularly those involving signal‐regulated transcription factors, remain poorly understood. Here, we investigated trans‐regulation of cytochrome P450 genes associated with sulfoxaflor resistance in Aphis gossypii Glover. Our findings suggest that G protein‐coupled receptor (GPCR)‐associated signaling and cAMP response element binding protein contribute to resistance‐associated P450 gene regulation. FISH and RT‐qPCR demonstrate that DopR1 transcripts are enriched in the epidermis, and β‐1A transcripts are enriched in the head. Molecular docking, heterologous expression, and biolayer interferometry assays demonstrate that DopR1 and β‐1A participate in cellular recognition or response to sulfoxaflor and activate downstream cAMP signaling. Silencing DopR1 and β‐1A reduced cAMP content, PKA activity, P450 enzyme activity, and resistance‐related P450 expression, thereby enhancing aphid sensitivity to sulfoxaflor. Inhibitor‐based suppression, pathway gene silencing, and dual‐luciferase reporter assays implicated AC‐2, AC‐5, PKA‐4, PKA‐7, CRTC, and CREB in regulating critical resistance‐related genes, especially CYP6CY19. Furthermore, functional validation using transgenic Drosophila melanogaster expressing CYP6CY13 and CYP6CY19 demonstrated increased sulfoxaflor tolerance. Collectively, these findings support a working model linking GPCR‐associated signaling to the regulation of selected resistance‐associated P450 genes in A. gossypii.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:09b0583a1ddf2000d8f77350495ea4fd8d480cff","kind":"journals","source":"Discover Food","title":"Harnessing the bioactive potential of Kaatha jamun (Syzygium cumini (L.) Skeels) through LC-MS/MS characterization of tissue-specific polyphenols with antioxidant and α-amylase inhibitory activities","url":"https://doi.org/10.1007/s44187-026-01348-3","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1007%2Fs44187-026-01348-3","date":"2026-11-15T00:00:00Z","timestamp":1794700800,"categories":["Docking & Screening"],"topic_ids":["docking"],"keywords":["molecular docking","bioactivity"],"matched_keywords":["molecular docking","molecular","compounds","bioactivity"],"matched_tags":["docking","generic_chem","bioactivity"],"doi":"10.1007/s44187-026-01348-3","external_id":"09b0583a1ddf2000d8f77350495ea4fd8d480cff","pdf_url":null,"code_url":null,"code_host":null,"authors":["Ajay Kumar","K. Singh","J. Prakash","A. K. Goswami","V. B. Patel","A. Singh","G. Mishra","Rajesh Kumar","Y. Yadav","A. Ingole","Bhupendra Sagore","Michi Tani","Vivek Saurabh","M. Kumar","Vishal Kumar"],"journal":"Discover Food","publisher":null,"impact_factor":null,"abstract":"Kaatha jamun ( Syzygium cumini (L.) Skeels), an underutilized small-fruited morphotype of jamun, represents a promising source of bioactive polyphenols with potential applications in functional foods and plant-based nutraceuticals. This study aimed to characterize the tissue-specific polyphenolic composition of Kaatha jamun (pulp, seed, and leaf) using LC–MS/MS and to evaluate its antioxidant and α-amylase inhibitory activities through integrated bioactivity assays, correlation analysis, and molecular docking. LC–MS/MS identified and quantified 33 polyphenols, including phenolic acids, flavonoids, and anthocyanins, revealing distinct tissue-specific distribution patterns. The leaf contained the highest sum of quantified phenolic acids. In contrast, the seed was enriched in flavan-3-ols, particularly catechin and epicatechin, as well as the flavonoid luteolin and the phenolic acids ellagic acid and sinapic acid. The pulp showed the highest sum of quantified anthocyanins, with delphinidin and cyanidin being the predominant compounds. The seed exhibited the strongest antioxidant activity (FRAP: 64.64 ± 2.99 mg TE/g; ABTS: 35.84 ± 3.24 mg AAE/g) and the highest α-amylase inhibitory activity (94.30%; IC₅₀: 37.56 µg/mL). Correlation analysis identified associations between selected phytochemicals and α-amylase inhibitory activity, including sinapic acid, ellagic acid, catechin, epicatechin, and kaempferol. Molecular docking predicted favourable interactions of luteolin and eriodictyol with the α-amylase active-site region. The observed tissue-specific differences in polyphenolic composition were associated with variation in antioxidant and α-amylase inhibitory activities. These findings highlight the diverse bioactive phytochemical composition of Kaatha jamun and provide a basis for further investigation of its potential as a source of functional food ingredients, while recognizing that the present evidence is based on in vitro assays, exploratory correlations, and predictive molecular docking.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:becfe45380e0d093b70300be045c22852f7ed5db","kind":"journals","source":"Journal of Heterocyclic Chemistry","title":"Highly Regio‐ and Diastereoselective Synthesis of Functionalized Tetrahydroquinolines via (4 + 2) Annulations of\n MBH\n Carbonates With 1‐Heterodienes","url":"https://doi.org/10.1002/jhet.70263","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1002%2Fjhet.70263","date":"2026-11-15T00:00:00Z","timestamp":1794700800,"categories":[],"topic_ids":[],"keywords":["molecular docking"],"matched_keywords":["molecular docking","molecular"],"matched_tags":["docking","generic_chem"],"doi":"10.1002/jhet.70263","external_id":"becfe45380e0d093b70300be045c22852f7ed5db","pdf_url":null,"code_url":null,"code_host":null,"authors":["Kai-Kai Wang","Xiao-Ran Wang","Q. Li","Quan-Xun Fang","Xia-Shi Gao","Xiao-Ran Wang","Rong-Xiang Chen","Guo-Yi Yan"],"journal":"Journal of Heterocyclic Chemistry","publisher":null,"impact_factor":null,"abstract":"A highly diastereo‐ and regioselective (4 + 2) annulation between o ‐acylamino‐aryl MBH carbonate and 1‐heterodienes has been established, enabling efficient construction of a broad library of functionalized tetrahydroquinolines bearing two vicinal tertiary/quaternary stereogenic centers. This transformation proceeds under mild conditions, delivering target products in high yields (up to 95%) with exclusive diastereocontrol (in all cases > 25:1 dr). Moreover, these new tetrahydroquinolines show promising potential as RORγ inhibitors in preliminary molecular docking studies. The relative configuration of a representative product in this series was unambiguously confirmed by single‐crystal x‐ray diffraction analysis.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:fe442b894de206cef08963327546ed01809e93f7","kind":"journals","source":"Universal Journal of Pharmaceutical Research","title":"IN SILICO EVALUATION OF THE INHIBITORY POWER OF AMIDINOUREA DERIVATIVES FOR PHOSPHOPANTETHEINYL TRANSFERASE IN MYCOBACTERIUM TUBERCULOSIS","url":"https://doi.org/10.22270/ujpr.v11i5.1665","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.22270%2Fujpr.v11i5.1665","date":"2026-11-15T00:00:00Z","timestamp":1794700800,"categories":["Property Prediction","Docking & Screening","ADMET & Safety","Free Energy & MD"],"topic_ids":["property","docking","admet","freeenergy"],"keywords":["molecular docking","QSAR","pharmacokinetic","molecular dynamics","enzyme"],"matched_keywords":["molecular docking","quantitative structure-activity relationship","QSAR","pharmacokinetic","drugs","molecular","chemical","molecular dynamics","enzyme"],"matched_tags":["docking","qsar_qspr","admet","generic_chem","molecular_dynamics","biology"],"doi":"10.22270/ujpr.v11i5.1665","external_id":"fe442b894de206cef08963327546ed01809e93f7","pdf_url":null,"code_url":null,"code_host":null,"authors":["Kouakou Jean-Louis Kouakou","Konan Auriol Kra","Elvice Akori Esmel","Mélalie Kéïtaépouse Guego","E. Megnassan"],"journal":"Universal Journal of Pharmaceutical Research","publisher":null,"impact_factor":null,"abstract":"Background: Tuberculosis (TB) is an infectious, contagious, and fatal disease, according to the 2025 Global Tuberculosis Report of the World Health Organization (WHO). Phosphopantetheinyl transferase (PptT) is an enzyme that acts as a catalyst in the biosynthesis of cellular lipids in Mycobacterium tuberculosis. It is considered potential target for the development of anti-tuberculosis drugs. Methods: We used molecular docking to establish a quantitative structure-activity relationship (QSAR) model for 25 amidinourea derivatives inhibitors of PptT. The binding free energies, a key physical descriptor, enabled us to elucidate the thermodynamic factors governing the interaction between the enzyme and the inhibitor. This approach, which combines structure-based molecular docking with quantitative modeling, identified specific, critical chemical substituents. This, in turn, led to the development of new derivatives that exhibit potent antitubercular activity. Results: The resulting QSAR model that accounts for 87% of the variation in biological activity, based on the binding free energies (BFE/PBSA) of the analogues used. Subsequently, an intuitive fragment-replacement approach, based on per-residue interaction energy, which indicates stabilizing interactions, was implemented to identify new analogues. Moreover, 200-ns molecular dynamics simulations have evaluated with satisfaction, the temporal evolution of various properties of the bound inhibitors, such as root mean square deviation (RMSD) from the initial conformation, radius of gyration (rGyr), number of intramolecular hydrogen bonds (intraHB), molecular surface area (molSA), solvent-accessible surface area (SASA), and polar surface area (PSA). The BFE/GBSA of the three best analogs, calculated over the last 50 nanoseconds of the trajectory, align with BFE/PBSA values. Finally, favorable values ​​for the pharmacokinetic descriptors of the best analogues are obtained using the QikProp, pkCSM, and mycoCSM programs. Conclusion: This study enabled us to identify a new amidinourea analogue AU_1-3-7-9, which is 10 times more active. Peer Review History: Received 22 July 2026; Reviewed 3 September 2026; Accepted 8 October; Available online 15 November 2026 Academic Editor: Prof. Dr. Gorkem Dulger, Duzce University, Turkey, gorkemdulger@yandex.com Reviewers: Dr. Sarfaraz Ahmed, Global Institute of Pharmaceutical Education and Research, Kashipur, Uttarakhand, India, sarfarazahmed1@gmail.com Sheima Nadim Kadhim, Deparment of Pharmacology and Toxicology, College of pharmacy, University of Basra, Iraq. ph.sheimanadim@yahoo.com","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:62311af1dc2506f247119603dbf9f93e9034855e","kind":"journals","source":"Journal of Drug Delivery and Therapeutics","title":"Integrating qNMR, Spectroscopic Characterization, and In Silico Screening for Comprehensive Evaluation of Curcumin and Its Drug-Like Derivatives","url":"https://doi.org/10.22270/jddt.v16i11.8066","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.22270%2Fjddt.v16i11.8066","date":"2026-11-15T00:00:00Z","timestamp":1794700800,"categories":[],"topic_ids":[],"keywords":["drug likeness","In Silico Screening","computational screening"],"matched_keywords":["drug-likeness","In Silico Screening","computational screening","Drug","molecular","Compounds","compound","protein"],"matched_tags":["drug_likeness","virtual_screening","generic_chem","biology"],"doi":"10.22270/jddt.v16i11.8066","external_id":"62311af1dc2506f247119603dbf9f93e9034855e","pdf_url":null,"code_url":null,"code_host":null,"authors":["P. Palak","Sneha Sneha","Parul Kumar","Ritika Gupta","Praveen Kumar","I. Kumar","K. Verma"],"journal":"Journal of Drug Delivery and Therapeutics","publisher":null,"impact_factor":null,"abstract":"Curcumin is an important natural polyphenol, but its quality and composition can vary depending on the source, and its poor physicochemical properties limit its pharmaceutical development. This study aimed to compare the quality of curcumin obtained from homemade turmeric from Indora, Kangra, Himachal Pradesh, India, and a locally marketed turmeric sample, while also characterizing the extracts and evaluating newly designed curcumin derivatives using computational methods. The turmeric samples were extracted using ethanol-based Soxhlet extraction and analyzed by ¹H NMR, FTIR, and mass spectrometry. Curcumin-related signals were observed in both extracts, with stronger signals in the marketed turmeric sample. Since complete internal-standard and validation data were not available, the NMR comparison was considered semi-quantitative. Mass spectrometry confirmed a protonated molecular ion at m/z 369.3 with characteristic fragments near m/z 177 and 149, while FTIR confirmed the major functional groups of curcumin. Fourteen curcumin derivatives were further evaluated using SwissADME and other in silico tools. Compounds 2, 4, 9, and 12 showed the most favourable overall drug-likeness profiles, although compound 12 showed low predicted gastrointestinal absorption. Compounds 2–7 and 9–11 showed high predicted gastrointestinal absorption, and target prediction suggested possible interactions with kinases, proteases, membrane receptors, and other protein classes. Overall, the study demonstrates a simple integrated approach for turmeric quality assessment, curcumin characterization, and computational screening of curcumin derivatives. Compounds 2, 4, and 9 were identified as promising candidates for further experimental studies, while compound 12 requires structural optimization. Further synthesis, validated qNMR analysis, and biological studies are needed to confirm these findings. Keywords: Curcuma longa, qNMR, FTIR, mass spectrometry, SwissADME, drug-likeness, BOILED-Egg, target prediction, natural-product drug discovery.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:ce84fc8a6a239652b916c1508f5d0f0f7809ab33","kind":"journals","source":"Food chemistry","title":"N-Succinylation of L-methionine broadens its taste-modulating profile: Enzymatic synthesis, sensory evaluation, soy sauce matrix assessment, and computational analyses","url":"https://doi.org/10.1016/j.foodchem.2026.152180","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1016%2Fj.foodchem.2026.152180","date":"2026-11-15T00:00:00Z","timestamp":1794700800,"categories":["Docking & Screening"],"topic_ids":["docking"],"keywords":["Molecular docking","density functional theory","molecular dynamics","receptor"],"matched_keywords":["Molecular docking","density functional theory","compound","compounds","Molecular","molecular dynamics","receptor"],"matched_tags":["docking","dft","generic_chem","molecular_dynamics","biology"],"doi":"10.1016/j.foodchem.2026.152180","external_id":"ce84fc8a6a239652b916c1508f5d0f0f7809ab33","pdf_url":null,"code_url":null,"code_host":null,"authors":["Pi-Miao Huang","Lu-Qi Qin","Zi-Jian Liang","Feng-Yuan Tang","Chun Cui"],"journal":"Food chemistry","publisher":null,"impact_factor":null,"abstract":"Reducing sodium chloride and monosodium glutamate without compromising palatability remains a challenge. This study examined whether succinylation could transform L-methionine, a reported umami enhancer, into a compound capable of enhancing both umami and saltiness. N-succinyl-L-methionine was prepared from L-methionine and succinic acid using food-grade enzymes, among which AY 50C gave the highest yield (43.3%). At 3 mg/L, neither L-Met nor N-Suc-Met elicited detectable umami or saltiness in water. In monosodium glutamate solutions, both compounds increased umami perception, with N-succinyl-L-methionine showing greater efficacy. In sodium chloride solutions, saltiness enhancement was observed only for N-succinyl-L-methionine. Sensory evaluation and electronic tongue analysis further supported stronger umami and saltiness responses induced by N-succinyl-L-methionine. Its taste-enhancing effect was also maintained in a diluted soy sauce matrix. Molecular docking, molecular dynamics simulations, and density functional theory indicated more favorable predicted receptor compatibility and electronic characteristics for N-succinyl-L-methionine than for L-methionine, which were consistent with the stronger sensory responses of N-succinyl-L-methionine. N-succinylation of L-methionine was associated with stronger umami enhancement and an additional saltiness-enhancing effect compared with L-methionine.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:c3a47c3e5da2a03fab3435ee6ed59312f8daa2db","kind":"journals","source":"International Journal of Molecular Sciences","title":"Network Toxicology and Machine Learning Prioritise a Seven-Gene Signature of Liver Injury Associated with Particle-Bound Polycyclic Aromatic Hydrocarbons","url":"https://doi.org/10.3390/ijms272210237","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.3390%2Fijms272210237","date":"2026-11-15T00:00:00Z","timestamp":1794700800,"categories":["Docking & Screening"],"topic_ids":["docking"],"keywords":["molecular docking","random forest"],"matched_keywords":["molecular docking","random forest","Machine Learning","molecular","chemical"],"matched_tags":["docking","ensemble","machine_learning","generic_chem"],"doi":"10.3390/ijms272210237","external_id":"c3a47c3e5da2a03fab3435ee6ed59312f8daa2db","pdf_url":null,"code_url":null,"code_host":null,"authors":["Jia-Chen Liang","Rui-Feng Yan"],"journal":"International Journal of Molecular Sciences","publisher":null,"impact_factor":null,"abstract":"As a major toxic component of environmental fine particulate matter (PM2.5), polycyclic aromatic hydrocarbons (PAHs) pose a significant risk to human health, particularly for liver damage, following inhalation or ingestion. However, their comprehensive molecular mechanisms underlying this hepatotoxicity have long remained unclear. In this study, we developed an integrative framework that combines network toxicology, batch-corrected two-stage machine learning, gene set variation analysis, immune-signature analysis and molecular docking. We retrieved 479 overlapping candidates from 570 PAH-related targets and 12,072 liver-injury-associated genes; the intersection of network hubs (MCC top-30 and MCODE modules) with differentially expressed genes across three merged hepatic cohorts (n = 174) yielded 32 candidates, from which three complementary algorithms (LASSO, SVM-RFE and permutation-calibrated random forest) distilled a seven-gene signature (CYP2C19, MYC, EGR1, HMMR, ICAM1, MAPK3 and NFE2L2). An eight-classifier comparison selected a support vector machine, exact Shapley values quantified each gene’s contribution, and a logistic risk score built on the seven genes discriminated metabolic dysfunction-associated steatotic liver disease (MASLD) with an area under the curve (AUC) of 0.953 in the training set, 0.828 in a held-out test set and 0.978 in an independent 216-biopsy validation cohort (10 controls and 206 MASLD cases), in which the score was discriminative but its probabilities were not well calibrated. GSVA and immune-signature analyses positioned the signature within a three-layer working model: collapse of metabolic and antioxidant capacity (CYP2C19, NFE2L2), imbalance of stress signalling (elevated MAPK3 with attenuated MYC and EGR1), and cell-behaviour remodelling (HMMR-linked proliferation programmes and ICAM1-linked immune remodelling), consistent with a state of heightened susceptibility to chemical stressors such as PAHs. In summary, this study provides a prioritised, hypothesis-generating target set and a reusable, batch-corrected analytical workflow for environmentally associated liver injury, with experimental validation of the prioritised targets as the necessary next step.","source_metadata":{"source":"semantic_scholar"}},{"id":"preprints:10.26434/chemrxiv.15012211/v1","kind":"preprints","source":"ChemRxiv","title":"OlefinCat-LLM: Full-Text Extraction and Chemistry-Guarded Ranking of Single-Site Polyolefin Catalyst–Process Packages","url":"https://doi.org/10.26434/chemrxiv.15012211/v1","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.26434%2Fchemrxiv.15012211%2Fv1","date":"2026-11-15T00:00:00Z","timestamp":1794700800,"categories":["LLMs & Agents"],"topic_ids":["llm"],"keywords":["LLM"],"matched_keywords":["LLM","compound","ligand"],"matched_tags":["LLMs","generic_chem"],"doi":"10.26434/chemrxiv.15012211/v1","external_id":"10.26434/chemrxiv.15012211/v1","pdf_url":null,"code_url":null,"code_host":null,"authors":["Xu Huang","Xuezhen Chen","Jiankun Li","Yucai Cao"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"An industrial polyolefin grade, whether an EPDM tire compound or a solution-process polyolefin elastomer, is specified by melt index or Mooney viscosity, but it is made by a precatalyst package (metal center, ligand family, and activator role) and by a polymerization window of temperature T , pressure P , and Al/M ratio. Those fields sit in the nested experimental tables of papers and patents. An abstract usually names only a catalyst class and a product, so a grade request cannot be turned into a rankable catalyst–process row from the abstract alone, and recombining common literature labels creates activation packages that were never reported together. OlefinCat-LLM turns nested tables into reviewable metal–ligand–activator–process records. Experimental reports are collected by automated screening, the tables are expanded from the full text, and each shell is ranked only after a chemistry guard and a process window anchored on a reported experiment. The guard is required because a metalligand similarity cannot see the activator: V/VOCl3 keeps a cosine of 0.934 with documented alkylaluminum chlorides and with blocked MAO or borate. On an EPDM tire request the guard contracts 2160 proposed shells to 216 documented packages. The top-ranked shell on that request is a late-metal ethylene package. Inverse ranking on this platform supplies only a reference for reviewing packages against the industrial request.","source_metadata":{"date_field":"posted","keywords":[]}},{"id":"journals:02d961b44e7ef36781de3bd2ee1886375d25b124","kind":"journals","source":"Black Sea Journal of Engineering and Science","title":"Research on Aptamer Integrated Surface-Enhanced Raman Spectroscopy Platforms for Diagnosis: Bibliometric Analysis of Topics and Trends","url":"https://doi.org/10.34248/bsengineering.1917710","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.34248%2Fbsengineering.1917710","date":"2026-11-15T00:00:00Z","timestamp":1794700800,"categories":["Cheminformatics","Property Prediction"],"topic_ids":["cheminformatics","property"],"keywords":["molecular fingerprint"],"matched_keywords":["molecular fingerprint","molecular"],"matched_tags":["molecular_fingerprints","generic_chem"],"doi":"10.34248/bsengineering.1917710","external_id":"02d961b44e7ef36781de3bd2ee1886375d25b124","pdf_url":null,"code_url":null,"code_host":null,"authors":["O. Ugurlu","Hatice Erdem","G. Açıkgöz"],"journal":"Black Sea Journal of Engineering and Science","publisher":null,"impact_factor":null,"abstract":"Aptamer integrated SERS platforms have become a remarkable and effective method in the diagnosis of various diseases, including cancer. In addition to the advantages of SERS, such as its unique molecular fingerprint identification and ease of use, aptamers offer multiple detection capabilities and more accurate measurements in complex environments. In this study, a bibliometric analysis was conducted on research related to aptamer-integrated SERS platforms for disease diagnosis in order to evaluate developments in the field. Considering all exclusion criteria, a total of 33 publications were obtained with 198 authors contributing and articles published in 24 journals. The average number of citations per document was 16.67, while the international collaboration rate was calculated as 15.15%. China, as the most productive and most cited country, seems to have a significant impact in this field. Furthermore, Cao Xiaowei, one of the most published authors, is Chinese and has the highest h-index among the authors. The most cited article was also produced in collaboration with China. In conclusion, it can be said that China haa a significant influence in this field. Keyword analysis also shows increasing interest in studies on the use of aptamer-based SERS systems in the diagnosis of Alzheimer's disease. A bibliometric analysis of publications on aptamer integrated SERS diagnostic platforms is not currently available in the literature. This study will significantly contribute to the authors gaining different perspectives and establishing new collaboration networks, thus enabling them to develop more effective strategies for use in diagnostic applications.","source_metadata":{"source":"semantic_scholar"}},{"id":"preprints:10.26434/chemrxiv.15012231/v1","kind":"preprints","source":"ChemRxiv","title":"SACred: mining the Protein Data Bank for single-atom catalyst sites","url":"https://doi.org/10.26434/chemrxiv.15012231/v1","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.26434%2Fchemrxiv.15012231%2Fv1","date":"2026-11-15T00:00:00Z","timestamp":1794700800,"categories":[],"topic_ids":[],"keywords":["MACE","DFT","PBE","force field","enzyme"],"matched_keywords":["MACE","DFT","PBE","force field","ligand","Protein","enzyme"],"matched_tags":["mlip_tools","dft","force_fields","generic_chem","biology"],"doi":"10.26434/chemrxiv.15012231/v1","external_id":"10.26434/chemrxiv.15012231/v1","pdf_url":null,"code_url":null,"code_host":null,"authors":["Frederic Schubert","Anna Silvestrova"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"Single-atom catalysts (SACs) are built from a small set of sites, most often a metal held by four nitrogens, while enzymes run hydrogen evolution, O2 and CO2 reduction and N2 fixation on single metals in many different first coordination shells. Here we mine that diversity. SACred, an automated pipeline built and run with AI agents, reduces every transition-metal site of the Protein Data Bank (PDB) to its first coordination shell, turns the recurring shells into candidate SAC sites and screens them in a funnel: each shell is mapped onto a single or double vacancy in graphene, screened with GFN2-xTB calibrated per metal against DFT, refined with PBE-D3 DFT for hydrogen adsorption (∆GH), formation energy (Ef) and dissolution potential (Udiss), simulated at 1,500 K with a fine-tuned MACE force field and checked against the literature by agents. From 291,977 metal sites in 53,868 PDB entries the pipeline builds 1,155 graphene sites for xTB and 123 for DFT. We validate each stage. Extraction agrees with the PDB’s own metal–ligand annotation (precision 0.98, recall 0.94). The mapping keeps all donors for 98 % of the sites but distorts single-vacancy geometries. In a prospective test, sites picked by calibrated xTB reached |∆GH| ≤ 0.3 eV in DFT in 67 % of cases, against 39 % for sites picked from enzyme knowledge alone and 44 % for a random sample from the xTB window (p = 0.06 and 0.07), and the whole xTB stage cost about 400 core-hours, as much as four to five","source_metadata":{"date_field":"posted","keywords":[]}},{"id":"journals:c66196a92def4ce2d038b680c2d870455b9f84a7","kind":"journals","source":"Life sciences","title":"SH2B1 modulates hepatic lipid droplet homeostasis and is associated with HSC70/LAMP2A-related PLIN2 regulation in metabolic dysfunction-associated fatty liver disease","url":"https://doi.org/10.1016/j.lfs.2026.124780","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1016%2Fj.lfs.2026.124780","date":"2026-11-15T00:00:00Z","timestamp":1794700800,"categories":[],"topic_ids":[],"keywords":["Molecular docking"],"matched_keywords":["Molecular docking","Molecular","protein"],"matched_tags":["docking","generic_chem","biology"],"doi":"10.1016/j.lfs.2026.124780","external_id":"c66196a92def4ce2d038b680c2d870455b9f84a7","pdf_url":null,"code_url":null,"code_host":null,"authors":["Zu-Shun Zhuo","Ying-Ying Hu","Lin-Yao Hao","Jiao-Yang Huang","Long Zhu","Xue-Juan Lin","Shan-Shan Ding"],"journal":"Life sciences","publisher":null,"impact_factor":null,"abstract":"AIMS SH2B adaptor protein 1 (SH2B1) participates in metabolic regulation, but its role in hepatic lipid-droplet homeostasis remains unclear. We investigated its role and underlying mechanisms in metabolic dysfunction-associated fatty liver disease (MAFLD). MATERIALS AND METHODS Human and mouse GEO transcriptomic datasets and liver single-cell RNA-sequencing data were analyzed to characterize SH2B1 expression and identify lipid-metabolism-related candidates. SH2B1 knockdown and overexpression were performed in palmitate/oleate (P/O)-treated AML12 hepatocytes, and lysosome-associated membrane protein 2 A (LAMP2A) knockdown was combined with SH2B1 overexpression to assess its functional contribution. Molecular docking explored potential interactions, while reciprocal co-immunoprecipitation and immunofluorescence assessed protein association and colocalization, respectively. In vivo, mice were fed a high-fat diet (HFD) and subsequently underwent AAV-mediated hepatic Sh2b1 knockdown. KEY FINDINGS SH2B1 expression was reduced in human and murine fatty liver datasets and was predominantly detected in hepatocytes. Perilipin 2 (PLIN2) expression was inversely correlated with SH2B1 across GSE165855 samples. P/O exposure reduced SH2B1 expression and increased PLIN2 levels and lipid accumulation. SH2B1 knockdown increased PLIN2 abundance and lipid storage, whereas overexpression produced the opposite phenotype. Reciprocal co-immunoprecipitation supported an association between SH2B1 and heat shock cognate 70 (HSC70), and immunofluorescence demonstrated their colocalization. SH2B1 manipulation bidirectionally altered LAMP2A abundance. LAMP2A knockdown attenuated the reductions in PLIN2 abundance and lipid accumulation induced by SH2B1 overexpression. Hepatic Sh2b1 knockdown exacerbated HFD-induced hepatic lipid accumulation and liver injury, with lower LAMP2A and higher PLIN2 levels. SIGNIFICANCE SH2B1 regulates hepatic lipid-droplet homeostasis, with LAMP2A contributing to its effects on PLIN2 abundance and lipid accumulation.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:4de1ceb8b4930c94d286189a88134f435ef0fe33","kind":"journals","source":"Journal of Drug Delivery and Therapeutics","title":"Short-Timescale Molecular Dynamics and MM-GBSA Analysis of a Designed Acridone (ARB-373) in Complex with LSD1/KDM1A","url":"https://doi.org/10.22270/jddt.v16i11.8061","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.22270%2Fjddt.v16i11.8061","date":"2026-11-15T00:00:00Z","timestamp":1794700800,"categories":["Docking & Screening"],"topic_ids":["docking"],"keywords":["AutoDock Vina","Molecular Dynamics","receptor"],"matched_keywords":["AutoDock Vina","Molecular","ligand","Molecular Dynamics","protein","receptor"],"matched_tags":["docking_software","generic_chem","molecular_dynamics","biology"],"doi":"10.22270/jddt.v16i11.8061","external_id":"4de1ceb8b4930c94d286189a88134f435ef0fe33","pdf_url":null,"code_url":null,"code_host":null,"authors":["Dinakaran Srimathi","G. Tamilarasi","K. Gowtham"],"journal":"Journal of Drug Delivery and Therapeutics","publisher":null,"impact_factor":null,"abstract":"Objective: To examine, by all-atom molecular dynamics (MD) simulation, whether the docking-derived binding pose of ARB-373 a designed fluoro-cyano-benzyl acridone analogue of the natural product arborinine is retained over the accessible simulation window in the LSD1/KDM1A site and to obtain exploratory implicit-solvent (MM-GBSA) estimate of the interaction energy of the complex.Methods: The top-ranked ARB-373 pose (AutoDock Vina 1.2.7; PDB 6TUY, FAD retained) was parameterised with AMBER99SB-ILDN (protein), GAFF2/ACPYPE (ligand), solvated in TIP3P water, neutralised to 0.15 M NaCl. After energy minimisation and NVT/NPT equilibration, a 100 ps production simulation was run in GROMACS 2023 at 310 K , 1 bar (2 fs step, LINCS, PME, 1.0–1.2 nm cut-offs). Backbone, ligand RMSD, per-residue Cα RMSF, radius of gyration, SASA, intermolecular hydrogen bonds and minimum protein–ligand distance were analysed. The interaction energy was computed by single-trajectory MM-GBSA (Generalized Born, igb=5; 100 frames; no entropy term).Results: The simulated receptor was conformationally steady over the window (mean backbone RMSD 0.069 nm, plateau ≈0.09 nm; radius of gyration 1.667 ± 0.006 nm). Ligand RMSD tracked the backbone (≈0.07 nm) for the first ~48 ps, then rose to ≈0.22 nm remained elevated, indicating that the ligand moved away from its docked pose; whether this reflects a genuine repositioning or a reorientation of the flexible N-benzyl group was not resolved. The minimum ligand–protein atom distance held at 0.206 ± 0.011 nm (≈2.1 Å), indicating continuous close contact but not, by itself, catalytic-pocket occupancy. Intermolecular hydrogen bonding was intermittent (0–1 bonds, mostly 0) and was not persistent even during the early pose-retained phase, so the docking-predicted Tyr761 hydrogen bond did not establish. Single-trajectory MM-GBSA gave ΔG(bind) ≈ −18.6 ± 2.3 kcal/mol, averaged across pre- and post-transition geometries.Conclusion: Over 100 ps the simulated complex did not rapidly rearrange and the ligand did not exit, but the ligand did not remain in its docked pose and the averaged MM-GBSA value corresponds to no single state. The docked pose is therefore not supported as a structural hypothesis. These are preliminary, feasibility-level observations; extended, replicated simulation with positive, negative controls is required before any binding-mode, affinity or inhibition claim.","source_metadata":{"source":"semantic_scholar"}},{"id":"preprints:10.26434/chemrxiv.15012213/v1","kind":"preprints","source":"ChemRxiv","title":"Stackable Force Fields in GROMACS: Addons and Mixed 1–4 Scaling","url":"https://doi.org/10.26434/chemrxiv.15012213/v1","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.26434%2Fchemrxiv.15012213%2Fv1","date":"2026-11-15T00:00:00Z","timestamp":1794700800,"categories":[],"topic_ids":[],"keywords":["Force Fields","force field","AMBER"],"matched_keywords":["parameters, and residue, hydrogen, and terminus databases to a base force field","Force Fields","force field","AMBER"],"matched_tags":["force_field_parameterization","force_fields"],"doi":"10.26434/chemrxiv.15012213/v1","external_id":"10.26434/chemrxiv.15012213/v1","pdf_url":null,"code_url":null,"code_host":null,"authors":["Alexey Shvetsov"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"Hybrid biomolecular simulations routinely combine building blocks described by different force fields, yet the GROMACS topology format and its preprocessing tools have long assumed a single, monolithic force field whose 1–4 electrostatic and Lennard-Jones scaling factors are global to the whole system. We introduce force field addons: self-contained .ffa directories that contribute atom types, bonded parameters, and residue, hydrogen, and terminus databases to a base force field at gmx pdb2gmx run time, with explicit compatibility declarations, priority-based resolution of overlapping definitions, and conflict detection. Native mixed 1–4 scaling complements them: the new [ addon_scaling ] and [ pair_scaling ] directives attach scaling factors to atom-type groups and to the torsion parameter records that prescribe the 1–4 interactions, with explicit wildcard-patterned overrides resolved by declared priority and silent inconsistencies turned into fatal preprocessing errors. With case-sensitive residue databases and automated AMBER-to-GROMACS conversion, GLYCAM-based glycans become usable alongside amber19sb proteins as first-class force field components; resolved topologies and per-term energies cross-check exactly against Amber references.","source_metadata":{"date_field":"posted","keywords":[]}},{"id":"journals:158777314e8f7de171c5194bfc9a24b7ace53494","kind":"journals","source":"Black Sea Journal of Engineering and Science","title":"Synthesis, DFT Studies, and Molecular Docking Evaluation of a Nortopsentin-Inspired Bis-Indole Furazan Derivative Targeting GSK-3Β","url":"https://doi.org/10.34248/bsengineering.1980010","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.34248%2Fbsengineering.1980010","date":"2026-11-15T00:00:00Z","timestamp":1794700800,"categories":["Docking & Screening","ADMET & Safety"],"topic_ids":["docking","admet"],"keywords":["drug likeness","Molecular Docking","inhibition constant","DFT","Density functional theory","ADME","kinase","enzyme"],"matched_keywords":["drug-likeness","Molecular Docking","inhibition constant","DFT","Density functional theory","ADME","Molecular","drug","kinase","enzyme"],"matched_tags":["drug_likeness","docking","inhibition_constant","dft","admet","generic_chem","biology"],"doi":"10.34248/bsengineering.1980010","external_id":"158777314e8f7de171c5194bfc9a24b7ace53494","pdf_url":null,"code_url":null,"code_host":null,"authors":["F. Sevgi","Yavuz Ekincioğlu","E. Karakaş Sarıkaya"],"journal":"Black Sea Journal of Engineering and Science","publisher":null,"impact_factor":null,"abstract":"Glycogen synthase kinase-3β (GSK-3β) is a key enzyme involved in various biochemical pathways and is a significant therapeutic target for neurodegenerative and metabolic diseases. In this study, a novel 3,4-bis(2-methyl-1H-indol-3-yl)-1,2,5-oxadiazole derivative (4c), inspired by the nortopsentin class of marine alkaloids, was designed and synthesized. The target bis-indole furazan derivative (4c) was efficiently synthesized via microwave-assisted cyclization of its bis-indole dioxime precursor (3a) in aqueous NaOH, providing a clean and catalyst-free strategy aligned with green chemistry principles. Structural characterization was achieved by FT-IR, ¹H-NMR, ¹³C-NMR, and elemental analysis and further validated through theoretical spectral calculations. Density functional theory (DFT) calculations provided insights into the molecular geometry, vibrational frequencies, and frontier molecular orbital characteristics of 4c. Molecular docking simulations against GSK-3β (PDB ID: 1Q41) revealed a favorable binding energy (-8.38 kcal/mol) and predicted inhibition constant of 0.725 µM. Computational ADME profiling also revealed acceptable drug-likeness properties. While the present findings are based solely on computational analyses, they indicate that bis-indole furazan derivatives could be considered as potential candidates for future optimization and experimental studies targeting GSK-3β.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:34606bd6ffe30fadc3559f67e257c56255ec195a","kind":"journals","source":"Antioxidants","title":"Targeting TUBA3E-Associated AIF Nuclear Accumulation Alleviates Trophoblast Injury in Preeclampsia","url":"https://doi.org/10.3390/antiox15111478","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.3390%2Fantiox15111478","date":"2026-11-15T00:00:00Z","timestamp":1794700800,"categories":["Docking & Screening"],"topic_ids":["docking"],"keywords":["molecular docking","molecular dynamics"],"matched_keywords":["molecular docking","molecular","molecular dynamics","protein"],"matched_tags":["docking","generic_chem","molecular_dynamics","biology"],"doi":"10.3390/antiox15111478","external_id":"34606bd6ffe30fadc3559f67e257c56255ec195a","pdf_url":null,"code_url":null,"code_host":null,"authors":["Lu-Jian Sun","Zi-Ming Ge","Miao-Miao Chen","Bing-Di Chao","J. Xu","Dong-Ni Huang","Hong-Li Liu","Lu-Mei Xie","Xin-Yue Hu","Min Li","Xiao-Lan Zhao","Xin Luo"],"journal":"Antioxidants","publisher":null,"impact_factor":null,"abstract":"Preeclampsia (PE) is a pregnancy-specific disorder characterized by placental dysfunction and maternal multisystem injury, with limited placenta-targeted therapies. Parthanatos, a PARP1-dependent regulated cell death pathway involving PAR accumulation, mitochondrial dysfunction, and apoptosis-inducing factor (AIF) nuclear translocation, has been implicated in cellular stress responses, but its role in trophoblast injury and AIF redistribution in PE remains unclear. Here, we investigated parthanatos-associated alterations in human PE placentas, L-NAME-induced PE-like mice, and MNNG-stressed trophoblast models. Increased PAR accumulation and nuclear AIF enrichment were observed in PE placentas and experimental PE-associated models, accompanied by mitochondrial abnormalities and cellular dysfunction. PARP inhibition partially protected trophoblasts from MNNG-induced injury. Bioinformatic screening identified TUBA3E as a candidate AIF-associated protein, which was further investigated using molecular docking and molecular dynamics simulations. TUBA3E was upregulated in PE and PE-like placentas and associated with AIF in stressed trophoblasts, whereas its depletion reduced AIF nuclear accumulation and alleviated oxidative stress, mitochondrial dysfunction, and NAD+/ATP depletion. 3,5-Dimethoxybenzyl alcohol (3,5-DA) reduced the stress-induced TUBA3E–AIF co-immunoprecipitation signal and attenuated MNNG-induced trophoblast injury. In L-NAME-induced PE-like mice, 3,5-DA reduced systolic blood pressure, improved fetal weight, and alleviated placental and renal pathological alterations, accompanied by reduced placental PAR accumulation and nuclear AIF levels. Collectively, these findings support an association between TUBA3E and AIF nuclear accumulation during trophoblast injury and demonstrate the protective effects of 3,5-DA in experimental PE-like models, suggesting its potential as a therapeutic candidate for PE.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:10.1021/acs.jctc.6c01510","kind":"journals","source":"Journal of Chemical Theory and Computation","title":"Thermodynamics of Metal–Hydroxide Interactions","url":"https://doi.org/10.1021/acs.jctc.6c01510","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1021%2Facs.jctc.6c01510","date":"2026-11-15T00:00:00+00:00","timestamp":1794700800,"categories":["Free Energy & MD"],"topic_ids":["freeenergy"],"keywords":["thermodynamic integration","density functional theory","free energy","molecular dynamics"],"matched_keywords":["thermodynamic integration","density functional theory","free energy","ligand","molecular","chemical","molecular dynamics"],"matched_tags":["free_energy_calcs","dft","thermodynamics","generic_chem","molecular_dynamics"],"doi":"10.1021/acs.jctc.6c01510","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Subhamoy Bhowmik","Zhen Li","Kenneth M. Merz Jr"],"journal":"Journal of Chemical Theory and Computation","publisher":"ACS","impact_factor":null,"abstract":"Metal-bound hydroxide species play central roles in biological catalysis, ligand exchange, and metal-ion selectivity, but their accurate representation in classical molecular dynamics simulations remains challenging because of the strong electrostatic and ion-induced dipole interactions associated with hydroxide coordination. In this work, we develop 12–6–4 Lennard-Jones nonbonded parameters for metal–hydroxide interactions involving Mg(II), Ca(II), Mn(II), Fe(II), Co(II), Ni(II), Cu(II), Zn(II), and Cd(II) in both TIP3P and OPC water models. The hydroxide–water C4 parameter was first optimized to reproduce the experimental hydration free energy of hydroxide, followed by systematic optimization of metal–hydroxide C4 terms against experimental metal–hydroxide binding free energies derived from stability constants. Potential of mean force calculations were used for parameter refinement, while thermodynamic integration calculations provided an independent validation through closed thermodynamic cycles. The optimized parameters reproduce experimental binding affinities within chemical accuracy for both water models and correct the severe overbinding observed with conventional 12–6 and default 12–6–4 models. Structural validation against density functional theory calculations further shows that the modified parameters provide reasonable metal–oxygen distances, with the OPC water model generally giving closer agreement with quantum mechanical geometries. Analysis of the potential of mean force (PMF) profiles reveals ligand-exchange barriers and solvent-mediated second-shell intermediates during hydroxide dissociation. Overall, the resulting parameter set provides a classical framework for modeling metal–hydroxide interactions in aqueous and biological environments where both thermodynamic accuracy and structural realism are required.","source_metadata":{"short_name":"JCTC","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"preprints:10.26434/chemrxiv.15012228/v1","kind":"preprints","source":"ChemRxiv","title":"λproc: a compositional language for process description, with a case study in chemical synthesis","url":"https://doi.org/10.26434/chemrxiv.15012228/v1","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.26434%2Fchemrxiv.15012228%2Fv1","date":"2026-11-15T00:00:00Z","timestamp":1794700800,"categories":[],"topic_ids":[],"keywords":["transformer"],"matched_keywords":["machine learning","transformer","chemical","molecule","molecules"],"matched_tags":["machine_learning","transformer","generic_chem"],"doi":"10.26434/chemrxiv.15012228/v1","external_id":"10.26434/chemrxiv.15012228/v1","pdf_url":null,"code_url":null,"code_host":null,"authors":["Luc E. Brunet"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"Processes—cooking recipes, formulations, industrial flowsheets, chemical synthesis procedures—are recorded as prose or as diagrams. Neither form is machine-checkable, neither admits a canonical string, and neither makes the reuse of a sub-process explicit. We present λproc, a language in which a process is written as a composition of functions over streams: a unit operation is a function from streams to streams, a sub-process is a named lambda, and the process graph is not drawn but emerges from referential transparency—two textual mentions of the same expression denote the same physical node. The language is specified by a grammar of five productions, a canonical form that maps each process to exactly one string, and a validator that decides whether a document denotes a well-formed process. We argue that these three properties make λproc a useful target for machine learning, and we test the claim on a case study. Converting 1 808 937 reactions text-mined from US patents yields 700 149 distinct λproc documents and, notably, zero invalid ones: the converter either produces a well-formed process or rejects the reaction. Trained on this corpus, a decoder-only transformer of 2.73 M parameters reads a target molecule and writes a synthesis procedure that validates in 99.7 % of cases and recovers 24.5 % of the reactants of the real procedure on molecules it has never seen. It is measured against two trivial baselines, and the comparison is instructive: a nearest-neighbour lookup over 597 350 training molecules ties it on the conventional metric, and diagnosing that tie shows the metric itself to be dominated by ubiquitous reagents—a constant answer scores 0.133 on it. On a metric restricted to building blocks the lookup falls to 0.180 while the model holds 0.266. Three further results follow. The grammar saturates at the smallest model size tested, so capacity buys chemistry, not syntax. The validator makes rejection sampling exact: ten samples yield ten usable procedures, with an oracle of 0.422 against 0.260 for the model’s own ranking, but no reference-free feature we measured recovers more than 0.005 of that gap. And conditioning on an inventory of available materials triples reactant recall to 0.764, where the model selects rather than copies the shelf—but stress tests show it trusts the shelf: when the right reactants are absent it falls below the unconditioned model, and when near-miss analogues stand in for them it substitutes them inside documents that still validate. A pretrained 350 M-parameter language model, fine-tuned on the same corpus in five configurations, loses by a factor of three. All code, weights and an interactive application are released with this paper.","source_metadata":{"date_field":"posted","keywords":[]}},{"id":"preprints:2611.19658v1","kind":"preprints","source":"arXiv","title":"Mesoscale Neural Operators: Learning Nonlinear Constitutive Laws from Molecular Dynamics for Scalable Fracture Predictions","url":"https://arxiv.org/abs/2611.19658v1","detail_url":"/chemradar/article?u=https%3A%2F%2Farxiv.org%2Fabs%2F2611.19658v1","date":"2026-11-14T23:04:16Z","timestamp":1794697456,"categories":[],"topic_ids":[],"keywords":["Molecular Dynamics"],"matched_keywords":["Machine learning","Molecular","Molecular Dynamics"],"matched_tags":["machine_learning","generic_chem","molecular_dynamics"],"doi":null,"external_id":"2611.19658v1","pdf_url":"https://arxiv.org/pdf/2611.19658v1","code_url":"https://github.com/fishmoon1234/PeridynamicNeuralOperator","code_host":"GitHub","authors":["Yue Yu","Jihong Wang","Stewart Silling","Ziwei Yang"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"Molecular dynamics (MD) simulations can capture the nanoscale physics of fracture at the atomistic level, but their computational cost prohibits predictions at the length scales and time scales needed for many applications. Machine learning approaches, on the other hand, have been proposed as heuristic surrogates for material fracture, yet they cannot guarantee the physical consistency of the predicted response. To address this shortcoming, we propose the mesoscale neural operator (MNO): a data-driven nonlocal constitutive model that is learned directly from MD loading and displacement data. After training, it can be used in continuum fracture simulations on much larger domains and grid sizes. The key idea of the MNO is to exploit the built-in characteristic length of a nonlocal model, called the horizon, to travel across scales. Physical consistency across scales is obtained by building two structural ingredients into the MNO: a normalized kernel that renders the strain energy density invariant under horizon rescaling, and a rescaled critical bond stretch that keeps the energy release rate consistent across scales. Through tests on synthetic data, we show that the MNO and its learning workflow indeed discover the nonlocal continuum multiscale model in the form of the nonlocal kernel and the nonlinear constitutive relation. We then apply the framework to coarse-grained MD data and find that in a test problem, the learned model matches the MD strain energy density and critical energy release rate within 3%. It also predicts the peak transmitted forces within 6% in out-of-distribution tension tests. As a further demonstration, the model is applied to a plate with 4000 times the area of the training specimens and containing an initial defect. In this problem, the model reproduces the main features of dynamic brittle fracture, including branch and cascading branch networks.","source_metadata":{"categories":["cs.LG","cond-mat.mtrl-sci"],"code_status":"found","code_url":"https://github.com/fishmoon1234/PeridynamicNeuralOperator"}},{"id":"preprints:2611.19365v1","kind":"preprints","source":"arXiv","title":"Exploring Energy Landscapes for Functional Materials: Ga vacancies in $β$-Ga$_{2}$O$_{3}$","url":"https://arxiv.org/abs/2611.19365v1","detail_url":"/chemradar/article?u=https%3A%2F%2Farxiv.org%2Fabs%2F2611.19365v1","date":"2026-11-14T12:08:08Z","timestamp":1794658088,"categories":["ML Potentials"],"topic_ids":["potentials"],"keywords":["NequIP"],"matched_keywords":["NequIP"],"matched_tags":["mlip_tools"],"doi":null,"external_id":"2611.19365v1","pdf_url":"https://arxiv.org/pdf/2611.19365v1","code_url":null,"code_host":null,"authors":["Soon Woo Park","Philipp Pracht","Seán R. Kavanagh","David J. Wales"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"Energy landscape exploration requires extensive energy and force evaluations, making efficient access to interatomic potentials an important practical challenge. We present a persistent socket interface that connects the landscape exploration tools in the GMIN, OPTIM, and PATHSAMPLE programs to potentials available through the Atomic Simulation Environment (ASE). The persistent design keeps both the calculator and socket connection active throughout the search, avoiding repeated initialisation. We apply this framework to understand and predict the properties of Ga vacancies in $β$-Ga$_2$O$_3$ using two NequIP-based potentials, and construct kinetic transition networks describing complex vacancy configurations and migration pathways. Temperature-dependent analysis of the dominant paths reveals how the most probable migration sequence can change, with the mechanism governed by global properties of the network, rather than simply the highest energy barrier. Visualisation, mean first-passage times, and heat capacities enable us to connect the organisation of the landscape to relaxation time scales and thermodynamic observables. The interface extends energy landscape methods to ASE-compatible atomistic models, providing a general framework for combining global optimisation, transition state searches, kinetic network analysis, and thermodynamic characterisation with recent developments in machine-learned interatomic potentials.","source_metadata":{"categories":["cond-mat.mtrl-sci","physics.chem-ph"],"code_status":"absent"}},{"id":"preprints:2611.19321v1","kind":"preprints","source":"arXiv","title":"Generation of protein conformations using a simple inverted transformer","url":"https://arxiv.org/abs/2611.19321v1","detail_url":"/chemradar/article?u=https%3A%2F%2Farxiv.org%2Fabs%2F2611.19321v1","date":"2026-11-14T10:32:49Z","timestamp":1794652369,"categories":[],"topic_ids":[],"keywords":["transformer","Molecular dynamics"],"matched_keywords":["neural network","deep learning","transformer","Molecular","Molecular dynamics","protein"],"matched_tags":["neural_network","machine_learning","transformer","generic_chem","molecular_dynamics","biology"],"doi":null,"external_id":"2611.19321v1","pdf_url":"https://arxiv.org/pdf/2611.19321v1","code_url":"https://github.com/llpalese/MDiT","code_host":"GitHub","authors":["Luigi Leonardo Palese"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"Molecular dynamics is an extremely powerful approach for studying protein behavior, but it requires enormous computational resources to obtain an adequate sampling of possible conformations. Recently, thanks to successes in predicting three-dimensional protein structures, there has been an explosion of research exploring the opportunities offered by deep learning to address the problem of protein conformational landscape. In this work, an artificial intelligence system based on a variant of the Transformer architecture is introduced and applied to the problem of determining the ensemble of three-dimensional structures that a protein can access and their statistical weight. After being trained on a dataset of conformations of a particular protein obtained using classical all-atom molecular dynamics, this neural network is able to directly generate physically plausible protein conformations similar to those obtained both through molecular dynamics and experimentally. The ordered sequence of conformations obtained as output of the neural network rapidly converges towards the most populated regions of the protein's conformational landscape, suggesting that the system may have learned not only the protein's three-dimensional organization but also the thermodynamic weights of the various feasible conformations. These features, combined with its simplicity and low computational load, make this Transformer variant an ideal candidate for the development of systems capable of performing generative molecular dynamics.","source_metadata":{"categories":["q-bio.QM","cs.AI","q-bio.BM"],"code_status":"found","code_url":"https://github.com/llpalese/MDiT"}},{"id":"preprints:2611.19072v1","kind":"preprints","source":"arXiv","title":"Nahual: A Sequence Model for Language and Atoms","url":"https://arxiv.org/abs/2611.19072v1","detail_url":"/chemradar/article?u=https%3A%2F%2Farxiv.org%2Fabs%2F2611.19072v1","date":"2026-11-14T01:22:49Z","timestamp":1794619369,"categories":["Design de novo","LLMs & Agents"],"topic_ids":["denovo","llm"],"keywords":["molecule generation","Generative models","diffusion model","LLMs","transformer"],"matched_keywords":["molecule generation","Generative models","diffusion model","LLMs","transformer","molecules","molecular","molecule"],"matched_tags":["de_novo_design","generative_model","diffusion","LLMs","transformer","generic_chem"],"doi":null,"external_id":"2611.19072v1","pdf_url":"https://arxiv.org/pdf/2611.19072v1","code_url":"https://github.com/the-matter-lab/nahual","code_host":"GitHub","authors":["Austin H. Cheng","Marcel Müller","Andreas Burger","Yeonghun Kang","Tsz Wai Ko","Marta Skreta","Luka Mucko","Ella Miray Rajaonson","Cher Tian Ser","Jérôme F. Gonthier","Alex Zook","Varinia Bernales","Alán Aspuru-Guzik"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"Generative models play a rapidly growing role in the design and simulation of molecules. The most widely used applications of generative models are in large language models (LLMs), owing to the natural flexibility of language for specifying and answering queries. However, generalist LLMs do not effectively handle large sets of 3D coordinates, whereas specialist generative models of 3D molecules cannot reach the controllability afforded by language. To bridge these capabilities, we propose Nahual, a decoder-only autoregressive diffusion model for any sequence of text and 3D molecules. To enable expressive autoregression on continuous coordinates, we propose next-token diffusion, whereby a causal transformer denoises each token while attending to the clean tokens before it, without increasing context size. We train Nahual on a curated set of twelve 3D chemistry tasks ranging from microsolvation to molecular conformer search to adsorption on metal surfaces, with associated evaluation metrics and baselines. The result is a single set of end-to-end-trained weights that simultaneously performs all tasks and surpasses state-of-the-art models specialized for crystal structure prediction and property-conditional molecule generation. Our results demonstrate the viability of scaling generalist decoder-only models for native multimodal understanding and generation across language and chemistry. Code is available at https://github.com/the-matter-lab/nahual.","source_metadata":{"categories":["cs.LG","cs.AI","cs.CL","physics.chem-ph"],"code_url":"https://github.com/the-matter-lab/nahual","code_status":"found"}},{"id":"journals:2bccfa047ce9f63663f001734743e207c7a5c05d","kind":"journals","source":"Scientia Pharmaceutica","title":"A Comparative Computational Evaluation of Rationally Designed Compounds Targeting NDM-1 Combining Molecular Docking, Molecular Dynamics, and MM/GBSA Calculations","url":"https://doi.org/10.3390/scipharm94040112","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.3390%2Fscipharm94040112","date":"2026-11-14T00:00:00Z","timestamp":1794614400,"categories":["Docking & Screening"],"topic_ids":["docking"],"keywords":["Molecular Docking","Molecular Dynamics","enzyme"],"matched_keywords":["Molecular Docking","Compounds","Molecular","ligand","ligands","Molecular Dynamics","enzyme","protein"],"matched_tags":["docking","generic_chem","molecular_dynamics","biology"],"doi":"10.3390/scipharm94040112","external_id":"2bccfa047ce9f63663f001734743e207c7a5c05d","pdf_url":null,"code_url":null,"code_host":null,"authors":["Eduvan Valencia Cristancho","M. Galvis","R. Escobar","F. Mesa","Yeison Alberto Garcés-Gómez"],"journal":"Scientia Pharmaceutica","publisher":null,"impact_factor":null,"abstract":"The global dissemination of bacteria producing New Delhi metallo-β-lactamase (NDM-1) severely compromises the effectiveness of β-lactam antibiotics and restricts available therapeutic options. Owing to the structural and functional complexity of this zinc-dependent enzyme, the development of effective inhibitors remains challenging. The objective of this study was to conduct a comparative in silico evaluation of the structural, dynamic, and energetic behavior of four rationally designed compounds (M1, M2, P3, and P4), derived from previous structure-guided design efforts, in complex with NDM-1. An integrated computational protocol was applied, including molecular docking using AutoDock4Zn, molecular dynamics simulations performed with GROMACS 2024.4, and binding free-energy calculations using the MM/GBSA approach implemented in gmx_MMPBSA. Structural descriptors (RMSD and RMSF), hydrogen-bond patterns, solvent-accessible surface area, residue–ligand distances, structural superposition analyses, and energetic contributions were evaluated throughout the simulation trajectories. In addition, the NDM-1–P4 complex, which exhibited the most favorable average MM/GBSA binding free-energy estimate in the initial analysis, was subjected to an extended 200 ns molecular dynamics simulation. All evaluated complexes exhibited negative binding free-energy estimates, with average ΔG_bind values ranging from −12.08 to −16.30 kcal/mol. Molecular docking predicted initial ligand accommodation within the catalytic region of NDM-1; however, molecular dynamics analyses revealed that all ligands progressively moved away from their initial docking poses and adopted alternative spatial arrangements during the simulations. The extended 200 ns simulation was consistent with this behavior for P4, which also lost sustained proximity to the catalytic Zn2+ ions despite retaining a favorable MM/GBSA energy estimate. These findings indicate that favorable MM/GBSA estimates may coexist with protein–ligand configurations that do not correspond to persistent occupancy of the catalytic site. Overall, the results demonstrate the complementary roles of molecular docking, molecular dynamics simulations, structural trajectory analyses, and binding free-energy calculations in characterizing predicted ligand recognition and structural persistence. The findings indicate that docking-derived poses and favorable binding free-energy estimates should not be interpreted independently of the structural behavior observed during molecular dynamics simulations. Importantly, none of these computational predictions alone establishes inhibitory activity against NDM-1, which requires experimental validation.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:7fb10efae646668062049eec680009de5b6a9366","kind":"journals","source":"Bioresource technology","title":"A multimodal model integrating microscopic images and fractal dimension for characterizing the granulation process of aerobic granular sludge","url":"https://doi.org/10.1016/j.biortech.2026.136260","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1016%2Fj.biortech.2026.136260","date":"2026-11-14T00:00:00Z","timestamp":1794614400,"categories":[],"topic_ids":[],"keywords":[],"matched_keywords":["machine learning","chemical"],"matched_tags":["machine_learning","generic_chem"],"doi":"10.1016/j.biortech.2026.136260","external_id":"7fb10efae646668062049eec680009de5b6a9366","pdf_url":null,"code_url":null,"code_host":null,"authors":["Zhe Liu","Zi-Bo Zhao","Jie Lei","Jia-Xuan Wang","Li Sun","Jia-Wei Mu","Yong-Jun Liu","Rong Chen"],"journal":"Bioresource technology","publisher":null,"impact_factor":null,"abstract":"Reliable identification of aerobic granular sludge (AGS) granulation stages can support cultivation monitoring and operational adjustment. We collected 231 microscopic images and particle size measurements from three sequencing batch reactors operated at influent chemical oxygen demand concentrations of 250, 500, and 1000 mg/L. Growth, stabilization, and disintegration were labeled according to consecutive-day changes in median particle diameter (ΔD50). Agreement between expert consensus and threshold-based labels was 88.31% (Cohen's κ = 0.823). Fractal dimension (FD) was calculated by box counting at 20, 40, and 80 pixels to characterize two-dimensional spatial occupancy of sludge foreground. Conventional machine learning, numerical-feature, image-only, image-numerical fusion, and vision-language models were compared. Trained models were evaluated using leave-one-reactor-out cross-validation across five training runs, with an 80:20 training-validation split in the two remaining reactors. FD was higher and less variable during stabilization (1.644 ± 0.082) than during growth (1.587 ± 0.087) or disintegration (1.580 ± 0.114). Adding FD increased the macro-averaged F1 score of ResNet34 from 0.866 to 0.905 and reduced confusion between stabilization and disintegration. ResNet34FD-Gated achieved the highest mean accuracy (0.919 ± 0.012) and Macro-F1 (0.920 ± 0.012). Input permutation analysis identified microscopic images as the dominant information source for stage classification. Integrating FD with microscopic images and particle size improved stage identification under laboratory conditions, although external validation remains necessary.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:5016ce056926f67beadacb195ebe18393362dfa7","kind":"journals","source":"Scientia Pharmaceutica","title":"Beyond Inert Fillers: A Systematic Review of Machine Learning Applications in Pharmaceutical Excipient and Formulation Design","url":"https://doi.org/10.3390/scipharm94040111","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.3390%2Fscipharm94040111","date":"2026-11-14T00:00:00Z","timestamp":1794614400,"categories":[],"topic_ids":[],"keywords":[],"matched_keywords":["Machine Learning","ML","ML models","drug"],"matched_tags":["machine_learning","generic_chem"],"doi":"10.3390/scipharm94040111","external_id":"5016ce056926f67beadacb195ebe18393362dfa7","pdf_url":null,"code_url":null,"code_host":null,"authors":["M. Kannavou","Gregory Davrazos"],"journal":"Scientia Pharmaceutica","publisher":null,"impact_factor":null,"abstract":"Background: Pharmaceutical excipients are increasingly recognized as active determinants of formulation performance, manufacturability, stability, and safety. Artificial intelligence (AI) and machine learning (ML) are being explored to support excipient-related decisions. Methods: This systematic review was conducted and reported according to PRISMA 2020. Scopus, Web of Science and PubMed were searched in October 2026 for English-language, peer-reviewed original research published between January 2020 and July 2026 that applied trained ML models to excipient-related formulation problems. Results: Of 463 unique records, 116 full-text reports were assessed and 60 studies were included; several share datasets or platforms and therefore represent fewer independent evidence streams. The studies were organized into five application domains: drug–excipient compatibility; dosage-form design and optimization; advanced manufacturing and drug-delivery systems; biopharmaceutical formulation and stability; and excipient safety and biological activity. Most studies relied on retrospective internal validation; only 12 reported experimental, prospective or external validation that fully met our criterion. Conclusions: AI/ML may help narrow the experimental search space and support excipient decisions, but evidence of reduced time, material use or cost is rarely reported. Implementation is limited by heterogeneous datasets, class imbalance, incomplete reporting, restricted applicability domains, limited external validation, and interpretability. Prospective benchmarking, interoperable databases, and regulatory-grade validation are needed.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:10.1021/acs.jafc.6c10516","kind":"journals","source":"Journal of Agricultural and Food Chemistry","title":"Binding Mechanisms\nof Catechins and Procyanidins to\nα-Lactalbumin at Acidic pH","url":"https://doi.org/10.1021/acs.jafc.6c10516","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1021%2Facs.jafc.6c10516","date":"2026-11-14T00:00:00+00:00","timestamp":1794614400,"categories":["Docking & Screening"],"topic_ids":["docking"],"keywords":["Molecular docking"],"matched_keywords":["Molecular docking","Molecular","Protein"],"matched_tags":["docking","generic_chem","biology"],"doi":"10.1021/acs.jafc.6c10516","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Amadou Adama Coulibaly","Kevin Billet","Saïd Bouhallab","Aurélie Nicolas","Pascaline Hamon","Sophie Guilois-Dubois","Sylvain Guyot","Martine Morzel"],"journal":"Journal of Agricultural and Food Chemistry","publisher":"ACS","impact_factor":null,"abstract":"Protein–polyphenol interactions can influence food functionality and nutritional properties. This study investigated the interactions of whey protein α-lactalbumin (α-Lac) with six polyphenols: epicatechin (EC), epigallocatechin (EGC), epicatechin gallate (ECG), epigallocatechin gallate (EGCG), procyanidin B2 (PB2), and procyanidin C1 (PC1), at pH 4, relevant to food conditions such as fruit-based products and yogurts. Fluorescence spectroscopy evidenced quenching by ECG and EGCG, whereas EGC had a modest effect. Intrinsic fluorescence of EC, PB2, and PC1 precluded their data analysis. Isothermal titration calorimetry revealed exothermic heat changes for ECG and EGCG, low-amplitude signals for EC, PB2, and EGC, and a reproducible but unstable signal for PC1. Turbidity clearly increased with ECG, EGCG, and PC1. Molecular docking suggested hydrophobic interactions and hydrogen bonds as the main driving forces. Overall, compared to data at neutral pH, an acidic environment might not modify fundamentally the interaction types but seemingly rearranged binding sites.","source_metadata":{"short_name":"J. Agric. Food Chem.","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"journals:10.1093/bioinformatics/btag839","kind":"journals","source":"Bioinformatics","title":"BioLM-Score: Language-Prior Conditioned Probabilistic Geometric Potentials for Protein-Ligand Scoring","url":"https://doi.org/10.1093/bioinformatics/btag839","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1093%2Fbioinformatics%2Fbtag839","date":"2026-11-14T00:00:00+00:00","timestamp":1794614400,"categories":["Property Prediction"],"topic_ids":["property"],"keywords":["virtual screening"],"matched_keywords":["virtual screening","Ligand","molecular","Protein"],"matched_tags":["virtual_screening","generic_chem","biology"],"doi":"10.1093/bioinformatics/btag839","external_id":null,"pdf_url":null,"code_url":"https://github.com/SZU-ADDG/BioLM-Score","code_host":"GitHub","authors":["Zhangfan Yang","Baoyun Chen","Dong Xu","Jia Wang","Ruibin Bai","Zexuan Zhu","Junkai Ji"],"journal":"Bioinformatics","publisher":"Oxford","impact_factor":5.4,"abstract":"Motivation Protein-ligand scoring must address affinity prediction, pose selection, and virtual screening, but these objectives are not necessarily optimized together. Geometric likelihood models may generalize poorly when training-set structural coverage is limited. We test whether pretrained biomolecular representations improve scoring under a controlled geometric baseline. Results BioLM-Score integrates frozen protein and molecular language-model embeddings into a residue-atom distance-likelihood framework via gated graph convolutions. We evaluate geometry-only, jointly trained, and two-stage fine-tuned variants, using the jointly trained variant for the primary controlled comparisons. All three DEKOIS 2.0 screening endpoints improve over the matched geometric baseline with target-level paired support, but affinity scoring and ranking do not improve uniformly. Under a time-split protocol, scoring is significantly lower and ranking is also lower than the geometric baseline, while screening gains persist. On BDB2020+, the difference in affinity correlation between BioLM-Score and GenScore narrows after LP-PDBBind-compatible retraining. Shuffling protein embeddings at inference reduces the screening advantage. However, retrained shuffle controls recover much of the performance, so correct alignment is not established as necessary. BioLM-Score strengthens geometric scoring on screening tasks without retraining the language models. Availability and implementation Code, dataset-split identifiers, preprocessing and evaluation scripts, and language-model embedding-generation utilities are available at https://github.com/SZU-ADDG/BioLM-Score. An archival code snapshot and pretrained model weights are available at doi:10.5281/zenodo.21878818. Supplementary information Supplementary data are available at Bioinformatics online.","source_metadata":{"short_name":"Bioinformatics","source":"crossref","abstract_source":null,"abstract_quality":null,"code_url":"https://github.com/SZU-ADDG/BioLM-Score","code_status":"found"}},{"id":"journals:4626d073d19dd02762b42dbd7e45b7a27f996db8","kind":"journals","source":"Asian Journal of Chemical Sciences","title":"Computational Elucidation of Putative Sperm-signalling Targets of Flueggea virosa: Network Pharmacology, Molecular Docking and Molecular Dynamics","url":"https://doi.org/10.9734/ajocs/2026/v16i6491","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.9734%2Fajocs%2F2026%2Fv16i6491","date":"2026-11-14T00:00:00Z","timestamp":1794614400,"categories":["Docking & Screening"],"topic_ids":["docking"],"keywords":["Molecular Docking","Molecular Dynamics"],"matched_keywords":["Molecular Docking","Molecular","ligand","Chemical","Molecular Dynamics","protein"],"matched_tags":["docking","generic_chem","molecular_dynamics","biology"],"doi":"10.9734/ajocs/2026/v16i6491","external_id":"4626d073d19dd02762b42dbd7e45b7a27f996db8","pdf_url":null,"code_url":null,"code_host":null,"authors":["Kehongo M. Nyanguru","A. Osano","Samwel Cheruiyot","E. Murungi","E. Kigondu"],"journal":"Asian Journal of Chemical Sciences","publisher":null,"impact_factor":null,"abstract":"Background: The methanolic root-bark fraction of Flueggea virosa (KBLM) has previously been associated with impaired human sperm function, but the signalling proteins potentially linked to its putatively annotated constituents remain uncertain. Aims: To identify putative sperm-signalling targets of constituents of the methanolic root-bark fraction of Flueggea virosa, previously shown to disrupt human sperm function. Study Design: An LC-MS/MS-informed, hypothesis-generating computational study integrating network pharmacology and structure-based modelling. Methodology: Putatively annotated KBLM constituents were subjected to protein-target prediction. Predicted targets were intersected with a sperm-associated protein dataset and analysed using protein-protein interaction networks and pathway mapping. Prioritised ligand–protein complexes were evaluated by molecular docking, 100-ns molecular dynamics simulations and MM-GBSA analysis. Results: Target prediction identified 279 proteins, of which 168 overlapped with a sperm-associated dataset of 5,596 proteins. Network analysis highlighted calcium, cyclic-nucleotide and phosphorylation-related signalling pathways, with AKT1 showing the highest centrality. AKT1, EGFR, PTGS2 and SRC were prioritised for structure-based evaluation. Docking scores ranged from -10.488 to -7.253 kcal/mol; kaempferol 3-O-(4-O-galloyl)-β-D-glucopyranoside yielded the lowest score against AKT1. During 100-ns simulations, AKT1 and PTGS2 exhibited comparatively constrained protein trajectories, and catechin movement in the SRC complex broadly tracked the protein. Both EGFR trajectories showed pronounced conformational displacement, with protein Cα RMSD exceeding 10 Å. MM-GBSA analysis further prioritised the AKT1 complexes; the galloylated kaempferol glycoside yielded a late-window ΔGbind estimate of -118.18 ± 9.10 kcal/mol. Conclusion: AKT1 and SRC emerged as experimentally tractable candidate targets, whereas EGFR and PTGS2 remained more provisional hypotheses. Chemical authentication and target-specific assays in human sperm are needed to validate these computational predictions.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:77abb6b3b720ced01e96267fa4721729c613831b","kind":"journals","source":"Discover Chemistry","title":"Computational identification of quinolinol based estrogen receptor alpha antagonists as promising therapeutics for MCF 7 breast cancer","url":"https://doi.org/10.1007/s44371-026-01046-0","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1007%2Fs44371-026-01046-0","date":"2026-11-14T00:00:00Z","timestamp":1794614400,"categories":["Property Prediction","Docking & Screening","ADMET & Safety"],"topic_ids":["property","docking","admet"],"keywords":["pharmacophore modeling","molecular docking","QSAR","density functional theory","DFT","ADMET","molecular dynamics","MD simulations","receptor"],"matched_keywords":["pharmacophore modeling","molecular docking","QSAR","density functional theory","DFT","ADMET","molecular dynamics","MD simulations","drug","molecular","compounds","compound","ligands","ligand","drugs","receptor"],"matched_tags":["pharmacophore","docking","qsar_qspr","dft","admet","molecular_dynamics","generic_chem","biology"],"doi":"10.1007/s44371-026-01046-0","external_id":"77abb6b3b720ced01e96267fa4721729c613831b","pdf_url":null,"code_url":null,"code_host":null,"authors":["M. Garhewal","Divya Pujari","A. Bhattacharya","Ayush Tiwari","Aditya Mishra","Avantika Agrawal","Shekhar Verma","Achal Mishra","Y. Vaishnav"],"journal":"Discover Chemistry","publisher":null,"impact_factor":null,"abstract":"Breast cancer is a major cause of cancer-related death worldwide, with estrogen receptor alpha (ER-α) being an important therapeutic target for hormone-dependent MCF-7 breast cancer. This study aimed to use a computer-aided drug design (CADD) methodology to assess 22 quinolinol analogues as potential ER-α antagonists. The CADD methodology consisted of Gaussian field-based three-dimensional quantitative structure–activity relationship (3D-QSAR) analysis, pharmacophore modeling, prediction of ADMET properties, molecular docking, molecular dynamics simulation (MD), and density functional theory (DFT) calculations. 3D-QSAR model exhibited good predictive ability (R2 = 0.9516, Q2 = 0.5989, RMSE = 0.82, and Pearson's r = 0.845). The results of the pharmacophore analysis showed that the critical features for anticancer activity are an aromatic ring, hydrophobicity, and hydrogen bond acceptor. According to the ADMET evaluation, compounds 1, 7, 25, and 13 have the highest potential to be developed into drug-like substances, with compound 1 having best ADMET score of 6.348. Molecular docking study of the ligands with ER-α (PDB ID: 3ERT) demonstrated that ligands 7, 1 and 25 exhibited significantly better binding affinities (− 11.98, − 11.48, and − 11.04 kcal/mol, respectively) than 4-hydroxytamoxifen (− 8.26 kcal/mol). MD simulations showed that ligand 7 formed a considerably more compact and stable ER-α complex through consistent hydrogen bond interactions during 100 ns simulation. DFT analyses also indicated that ligand 7 had kinetic stability and significantly better electronic properties than the other ligands studied suggesting it for the development of breast cancer drugs that can target ER-α. A drug discovery workflow identifies ligand 7 as a candidate for breast cancer therapy, with in vitro and in vivo validation still required. Dataset preparation using field-based 3D QSAR and pharmacophore modeling is followed by ADMET screening, molecular docking, a 100-nanosecond molecular dynamics simulation, and density functional theory analysis of frontier molecular orbitals.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:10.1186/s13321-026-01327-x","kind":"journals","source":"Journal of Cheminformatics","title":"CTMP: bounded no-reference protein–ligand pose repair with failure-aware virtual-screening audit","url":"https://doi.org/10.1186/s13321-026-01327-x","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1186%2Fs13321-026-01327-x","date":"2026-11-14T00:00:00+00:00","timestamp":1794614400,"categories":["Cheminformatics"],"topic_ids":["cheminformatics"],"keywords":[],"matched_keywords":[],"matched_tags":["biology","generic_chem"],"doi":"10.1186/s13321-026-01327-x","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Yafei Huang","Yinggang Ling","Ya Xu","Shenkuang Wu","Wenping Feng"],"journal":"Journal of Cheminformatics","publisher":"Springer","impact_factor":5.7,"abstract":"","source_metadata":{"short_name":"J. Cheminform.","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"journals:7f7a90c6a28755acfb1401e88c04c113a150be34","kind":"journals","source":"Protein Science","title":"Designed multiple sequence alignments guide AlphaFold for improved modeling of loop‐mediated protein–protein interactions","url":"https://doi.org/10.1002/pro.70882","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1002%2Fpro.70882","date":"2026-11-14T00:00:00Z","timestamp":1794614400,"categories":[],"topic_ids":[],"keywords":["AlphaFold3"],"matched_keywords":["AlphaFold3","deep learning","molecular","protein"],"matched_tags":["ml_protein_ligand","machine_learning","generic_chem","biology"],"doi":"10.1002/pro.70882","external_id":"7f7a90c6a28755acfb1401e88c04c113a150be34","pdf_url":null,"code_url":null,"code_host":null,"authors":["Lourdes Carcelén","Alexandre Casadesús","E. Marcos"],"journal":"Protein Science : A Publication of the Protein Society","publisher":null,"impact_factor":null,"abstract":"Accurate modeling of synthetic protein–protein interactions is essential for designing therapeutics, diagnostics, and molecular imaging reagents. Although deep learning has transformed protein structure prediction, modeling heteromeric interactions mediated by flexible loops—as in antibodies, nanobodies, and monobodies—remains challenging due to weak coevolutionary signals at binding interfaces and substantial conformational variability. Here, we present a framework that guides AlphaFold predictions of loop‐mediated interactions using synthetic multiple sequence alignments (MSAs) generated through ProteinMPNN interface design. Benchmarks across nanobody–, antibody–, and monobody–antigen complexes show that designed MSAs encode interface‐specific contact information that can be leveraged in two complementary ways. First, AlphaFold2‐Multimer and AlphaFold3 predictions conditioned on designed MSAs produce confident and accurate structures for compatible interfaces, while yielding low‐confidence or structurally inconsistent predictions for incompatible sequences or docking poses, enabling improved evaluation of candidate protein–protein complex models—particularly relevant for protein design. Second, when applied to AlphaFold‐predicted structures, designed MSAs guide structural refinement and reranking, enabling the development of a strategy that improves blind structure prediction in low‐confidence cases without exhaustive sampling. Together, these results establish designed MSAs as controllable inputs for structure prediction, with applications in protein design and blind prediction of challenging protein–protein interactions.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:baa08b5af322d8be384c9d48c3d39000e8aeb031","kind":"journals","source":"Computer methods and programs in biomedicine","title":"DHGCN: A topology-driven multivariate entity association identification model for drug repositioning via decoupled hypergraph convolution","url":"https://doi.org/10.1016/j.cmpb.2026.109714","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1016%2Fj.cmpb.2026.109714","date":"2026-11-14T00:00:00Z","timestamp":1794614400,"categories":[],"topic_ids":[],"keywords":["graph neural networks","denoising","attention mechanism"],"matched_keywords":["neural networks","graph neural networks","denoising","attention mechanism","drug"],"matched_tags":["neural_network","gnn","diffusion","transformer","generic_chem"],"doi":"10.1016/j.cmpb.2026.109714","external_id":"baa08b5af322d8be384c9d48c3d39000e8aeb031","pdf_url":null,"code_url":null,"code_host":null,"authors":["Chen-Feng Zhang","Jian-Ping Zhao","Fei Wang","Jun-Feng Xia"],"journal":"Computer methods and programs in biomedicine","publisher":null,"impact_factor":null,"abstract":"Computational drug repositioning has emerged as a promising strategy to accelerate the drug development process for complex human diseases. However, accurately identifying higher-order associations among multimodal biomedical entities remains a significant computational challenge. While graph neural networks have been widely applied to biomedical network mining, existing approaches often struggle with semantic coupling and feature over-smoothing when processing heterogeneous relationships. To address these limitations, we propose the Decoupled Hypergraph Convolutional Network (DHGCN), a topology-driven framework designed for multivariate entity association identification. Specifically, DHGCN constructs independent hypergraph convolutional architectures to project parallel semantic subspaces tailored to specific interaction patterns. The structural configuration facilitates the topological decoupling of heterogeneous signals, thereby mitigating multimodal semantic conflicts. Furthermore, we introduce a strategy of independent decoupling followed by feature aggregation, alongside a dynamic denoising mechanism, to capture cross-modal complementarities and filter topological noise. A unified multi-view query attention mechanism is also deployed to adaptively aggregate high-value associative semantics. Experimental results on the large-scale Drug-Gene-Disease dataset demonstrate that DHGCN achieves the best accuracy, precision, recall and F1, while remaining competitive on AUPR and AUC in predicting multivariate entity associations. Ablation studies, sensitivity analyses, and interpretability analyses further validate the efficacy and robustness of its decoupling mechanism. Overall, our model offers a robust computational paradigm for heterogeneous network mining and holds great potential for advancing precision drug repositioning.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:10.1038/s41598-026-78991-4","kind":"journals","source":"Scientific Reports","title":"Discrimination of ranges of closely related RNA targets using CRISPR-based detection assay developed using machine learning-based optimization","url":"https://doi.org/10.1038/s41598-026-78991-4","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1038%2Fs41598-026-78991-4","date":"2026-11-14T00:00:00+00:00","timestamp":1794614400,"categories":[],"topic_ids":[],"keywords":[],"matched_keywords":["machine learning","molecular"],"matched_tags":["machine_learning","generic_chem"],"doi":"10.1038/s41598-026-78991-4","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Tomasz A. Leski","Scott N. Dean","Zachary T. Johnson","Christopher M. Green","Sebastián A. Díaz","David A. Stenger"],"journal":"Scientific Reports","publisher":"Nature","impact_factor":null,"abstract":"Early detection and identification of pathogen strains are crucial for mitigating infectious disease threats, including those posed by closely related zoonotic variants. A key limitation of current molecular detection approaches is their narrow focus on individual targets, which can delay recognition of outbreaks caused by closely related but non-identical pathogens. Here we report a two-channel CRISPR/Cas13a detection assay capable of discriminating between SARS-CoV-2 and a panel of twelve closely related zoonotic sarbecoviruses using a compact duplex assay format. Guide RNA sets with tunable specificity were designed using a two-stage machine learning pipeline: initial candidate filtering with a pre-existing RuleFit model, followed by retraining on 382 sarbecovirus-specific in vitro data points to refine the mismatch threshold from 4 to 6 mismatches for target exclusion. The optimized assay achieved proof-of-concept three-way classification of synthetic RNA targets into: (1) SARS-CoV-2, (2) the closest known near-neighbor BANAL-236, and (3) remaining zoonotic near-neighbors, with a limit of detection of 15–150 fM for synthetic RNA targets. This approach provides a practical framework for surveillance-oriented detection that identifies not only a specific pathogen but also its closest relatives, provided that system-specific validation and model retraining are included for new target systems.","source_metadata":{"short_name":"Sci. Rep.","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"journals:10.1038/s41467-026-79090-8","kind":"journals","source":"Nature Communications","title":"DPI-score: a deep learning-based metric for assessing protein-protein interfaces in cryoEM derived assemblies","url":"https://doi.org/10.1038/s41467-026-79090-8","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1038%2Fs41467-026-79090-8","date":"2026-11-14T00:00:00+00:00","timestamp":1794614400,"categories":[],"topic_ids":[],"keywords":[],"matched_keywords":["deep learning","protein"],"matched_tags":["machine_learning","biology"],"doi":"10.1038/s41467-026-79090-8","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Niraj Bhujel","Nick Whyatt","Agnel Praveen Joseph","Jyothish Thomas","Luc G. Elliot","Tamilarasan Sabapathy","Jeyan Thiyagalingam","Sony Malhotra"],"journal":"Nature Communications","publisher":"Nature","impact_factor":15.7,"abstract":"Advances in cryoEM have led to a surge in high-resolution structures; however, model building and refinement at resolutions ( > = 3 Å) and in regions with variable local resolution remain challenging, making validation essential. CryoEM derived assemblies often contain extensive protein-protein interfaces, however, most existing validation metrics focus on density fit or overall geometry without directly assessing interface quality. To address this, we present DPI-Score, a deep learning-based metric for assessing protein-protein interfaces in cryoEM derived complexes. The method uses only raw structural coordinates of interface atoms, without requiring engineered features, and achieves 87.53% validation accuracy. DPI-Score was applied to 29,120 interfaces from 6,011 fitted entries with resolutions worse than 3 Å in the Electron Microscopy Data Bank. Here, we show that DPI-Score provides complementary information to existing validation metrics and can identify interface errors in modelled assemblies that are not detected by density-based scores alone.","source_metadata":{"short_name":"Nat. Commun.","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"journals:7a585e7078c171b45dcc72a7f0050304e691f520","kind":"journals","source":"Foods","title":"Effects of Flavonol Treatment with Different B-Ring Substitution Patterns on the Structural and Functional Properties of Whey Protein Isolate","url":"https://doi.org/10.3390/foods15224082","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.3390%2Ffoods15224082","date":"2026-11-14T00:00:00Z","timestamp":1794614400,"categories":["Docking & Screening"],"topic_ids":["docking"],"keywords":["Molecular docking"],"matched_keywords":["Molecular docking","Molecular","Protein"],"matched_tags":["docking","generic_chem","biology"],"doi":"10.3390/foods15224082","external_id":"7a585e7078c171b45dcc72a7f0050304e691f520","pdf_url":null,"code_url":null,"code_host":null,"authors":["Zhuo Chen","Lin-Lin Xiong","Xin-Yu Wu","Shu-Hui Lv","Yue Liu","Xiao-Meng Sun","Zhi-Qing Zhang","Man Zhou","Qiang Cui"],"journal":"Foods","publisher":null,"impact_factor":null,"abstract":"The effect of differences in the position of substituents on the B ring of flavonols on the structural and functional properties of whey protein isolate (WPI) remains unclear. In this study, isorhamnetin (Iso), which contains a 3′-methoxy group and a 4′-hydroxyl group, and morin (Mor), which contains hydroxyl groups at the 2′ and 4′ positions, were used to treat WPI under the same alkaline conditions. The Folin–Ciocalteu-derived apparent flavonol equivalent of Iso-WPI was 113.49 mg/g, approximately 2.8 times that of Mor-WPI (41.10 mg/g). Molecular docking predicted distinct non-covalent interactions, and spectroscopic analyses suggested changes in aromatic-residue microenvironments, hydrogen-bonding environments, and protein secondary structure. Both flavonols reduced the relative ANS fluorescence intensity and particle size of WPI. Iso-WPI exhibited the lowest relative ANS fluorescence intensity and the highest emulsifying activity index, whereas Mor-WPI showed the smallest particle size and the highest apparent endothermic peak temperature. Mor-WPI also showed the highest foaming capacity, the largest total volume after 30 min in the foaming assay, the highest emulsion stability index, and the strongest DPPH radical-scavenging response. Overall, Iso-WPI showed the strongest initial emulsifying performance, whereas Mor-WPI performed best in the 30 min foaming, emulsion stability, and DPPH assays under the tested conditions.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:a434a1736ac3ed189f5fb7c8a52d4235cd547120","kind":"journals","source":"Animal reproduction science","title":"Fisetin mitigates cryodamage in buck spermatozoa by enhancing antioxidant capacity, preserving ultrastructure, and regulating ferroptosis-related pathways","url":"https://doi.org/10.1016/j.anireprosci.2026.108382","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1016%2Fj.anireprosci.2026.108382","date":"2026-11-14T00:00:00Z","timestamp":1794614400,"categories":[],"topic_ids":[],"keywords":["Molecular docking"],"matched_keywords":["Molecular docking","Molecular"],"matched_tags":["docking","generic_chem"],"doi":"10.1016/j.anireprosci.2026.108382","external_id":"a434a1736ac3ed189f5fb7c8a52d4235cd547120","pdf_url":null,"code_url":null,"code_host":null,"authors":["W. Khalil","A. A. Ismail","M. El-Harairy","Ahmed I. Yousif","M. Hassan","R. Sindi","Asmaa M. Sheiha","S. Abdelnour"],"journal":"Animal reproduction science","publisher":null,"impact_factor":null,"abstract":"Oxidative stress severely damages sperm during cryopreservation. Fisetin (FT), a flavonoid with strong antioxidant properties, has multifaceted functions, however, its protective potential during semen cryopreservation remains unexplored. This experiment evaluates the effects of supplementing the cryopreservation medium with various levels of FT on buck semen quality, focusing on post-thaw sperm variables, apoptosis, oxidative and antioxidant status, acrosomal integrity and ultrastructural, and reactive oxygen species (ROS) levels. Semen (60 ejaculates) was collected from five healthy bucks using an artificial vagina. Ejaculates meeting pre-established quality criteria were pooled to minimize individual male variability and subsequently diluted in a Tris-based extender. The control group (FT0) received no additive, while the treatment groups were supplemented with FT at concentrations of 6.25, 12.5, 25, and 50 µM (designated as FT6.25, FT12.5, FT25, and FT50, respectively). Statistical analysis was performed using one-way ANOVA. Fisetin (FT) supplementation at 12.5, 25, and 50 µM significantly enhanced progressive motility and viability compared to the control group (P < 0.01) following equilibration at 5°C for 4 h. Compared with the FT0 control, the FT50 group showed the highest functional membrane integrity alongside significantly improved post-thaw kinematic parameters, including DAP, DSL, VSL, and STR (P < 0.05). Following a 2-hour incubation post-thaw (37 °C, 5% CO2), all FT-treated groups maintained higher sperm motility and viability than the control, with the greatest membrane integrity observed in the FT25 and FT50 groups (P < 0.05). Furthermore, FT addition significantly increased the proportion of live, acrosome-intact sperm while reducing dead, acrosome-intact cells (P < 0.05). Antioxidant activities (CAT and TAC) were significantly elevated in FT-treated groups, peaking in the FT50 group (P < 0.05). Conversely, intracellular ROS and markers of oxidative stress (MDA and H2O2) decreased significantly across all FT groups (P < 0.001), with nitric oxide (NO) reaching its lowest concentration at 50 µM FT (P = 0.003). Furthermore, the FT25 and FT50 groups exhibited maximum sperm viability alongside the lowest apoptosis rates (P < 0.001). Transmission electron microscopy (TEM) confirmed that FT particularly at 25 and 50 µM preserved sperm ultrastructure, as evidenced by intact nuclei, normal mitochondrial architecture, and structurally sound plasma membranes. Molecular docking suggested possible binding energies between FT and the sperm ferroptosis-related proteins ferritin, GPX4, and VDAC2 of -4.51, -4.27, and -5.74 kcal/mol, respectively. In conclusion, fisetin mitigates cryodamage through potent antioxidant and anti-apoptotic pathways, significantly improving post-thaw semen quality. Supplementing cryopreservation extenders with fisetin (specifically at 25-50 µM) represents a promising strategy to optimize artificial insemination and reproductive outcomes in the livestock industry.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:10.1021/acs.jcim.6c02076","kind":"journals","source":"Journal of Chemical Information and Modeling","title":"From Molecular Structure to Human Pharmacokinetics:\nClearance Prediction Drives Exposure Error in a Hybrid ML–PBPK\nWorkflow","url":"https://doi.org/10.1021/acs.jcim.6c02076","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1021%2Facs.jcim.6c02076","date":"2026-11-14T00:00:00+00:00","timestamp":1794614400,"categories":["Cheminformatics","Property Prediction","ADMET & Safety"],"topic_ids":["cheminformatics","property","admet"],"keywords":["Clearance Prediction","pharmacokinetic","PK","molecular representation"],"matched_keywords":["ML","Clearance Prediction","pharmacokinetic","PK","Molecular","compound","compounds","molecular representation"],"matched_tags":["machine_learning","metabolism","admet","generic_chem","molecular_descriptor"],"doi":"10.1021/acs.jcim.6c02076","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Lianjin Cai","Amitesh Badkul","Jingchen Zhai","Mai McWilliams","Lei Xie","Junmei Wang"],"journal":"Journal of Chemical Information and Modeling","publisher":"ACS","impact_factor":5.3,"abstract":"End-to-end structure-to-exposure workflows combine molecular representation, pharmacokinetic (PK) input prediction, and physiologically based pharmacokinetic (PBPK) simulation, but aggregate accuracy does not reveal where error enters the workflow. We developed graph-based models for predicting plasma unbound fraction, systemic clearance, and steady-state volume of distribution and used paired input substitution in fixed compound-specific PBPK models to trace the downstream effects of these predictions. On held-out benchmark sets, models combining graph-derived and physicochemical features performed comparably to descriptor-based controls, with no detectable paired differences. In 41 compounds, replacing observed clearance with predicted intrinsic or systemic clearance caused the largest deterioration in exposure prediction. AUC0–t 2-fold coverage decreased from 82.9% to between 41.5 and 53.7%, whereas substitutions of physicochemical properties, plasma unbound fraction, and volume of distribution had smaller effects. The intrinsic-clearance and systemic-clearance workflows were not detectably different under matched input slots. Persistent Cmax underprediction with observed inputs indicated an additional limitation in distribution-model structure. When all three PK inputs were predicted, AUC0–t and Cmax were within 2-fold for 41 and 59% of compounds. Because encoder exposure may favor performance in this set, these coverage estimates are upper bounds for unseen chemistry. The paired-substitution design therefore identifies limited clearance prediction accuracy, rather than the PBPK clearance parametrization strategy, as the principal source of exposure error in the evaluated workflow and separates it from residual PBPK structural error.","source_metadata":{"short_name":"JCIM","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"journals:10.1021/acs.jctc.6c00944","kind":"journals","source":"Journal of Chemical Theory and Computation","title":"Functional and Density-Driven Errors in Density Functional\nTheory: Quantum Monte Carlo Benchmarks for Solids","url":"https://doi.org/10.1021/acs.jctc.6c00944","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1021%2Facs.jctc.6c00944","date":"2026-11-14T00:00:00+00:00","timestamp":1794614400,"categories":[],"topic_ids":[],"keywords":["machine learning potential","PBE"],"matched_keywords":["machine learning potential","machine learning","PBE"],"matched_tags":["mlip_tools","machine_learning","dft"],"doi":"10.1021/acs.jctc.6c00944","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Ayoub Aouina","Nicolas Tancogne-Dejean","Silvana Botti"],"journal":"Journal of Chemical Theory and Computation","publisher":"ACS","impact_factor":null,"abstract":"We introduce a systematic analysis of density functional approximation errors in solids by separating functional-driven from density-driven contributions using quantum Monte Carlo densities of silicon, sodium chloride, and copper as reference. Typically, functional errors dominate, but we identify important exceptions where the xc component of the density-driven error exceeds the functional error, notably for SOGGA11 in the semiconductor and the insulator. Material dependence is striking: 61% of functionals show error cancellation in silicon versus 16% in copper, and only five functionals surpass LDA accuracy for metallic copper even with exact densities. For silicon and sodium chloride, GILL or BECKE exchange combined with PBE, PW91, or P86 correlation achieves near-exact xc energies on QMC densities, while copper requires specialized functionals like PBEsol or PBELYP. High-quality densities consistently reduce density-driven errors across all systems. Historical analysis reveals that 1990s GGA functionals outperform many modern meta-GGAs, contradicting expectations of systematic improvement along Jacob’s ladder. These results provide practical guidance for functional selection and highlight implications for machine learning potential development, where material-dependent error cancellation may compromise transferability.","source_metadata":{"short_name":"JCTC","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"journals:10.1038/s41598-026-79150-5","kind":"journals","source":"Scientific Reports","title":"GC–MS metabolite profiling, elemental composition, and molecular docking analysis of camel milk extracts: insights into antioxidant and antibacterial activities","url":"https://doi.org/10.1038/s41598-026-79150-5","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1038%2Fs41598-026-79150-5","date":"2026-11-14T00:00:00+00:00","timestamp":1794614400,"categories":["Docking & Screening"],"topic_ids":["docking"],"keywords":["molecular docking"],"matched_keywords":["molecular docking","molecular"],"matched_tags":["docking","generic_chem"],"doi":"10.1038/s41598-026-79150-5","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Gashaw Nigussie","Sumera Zaib","Aman Dekebo","Markos Abebe","Metasebia Tegegn","Abebe Bitew Zegeye","Asfaw Meressa","Temesgen Negassa","Messay Woldemariam","Mo Hunsen","Milkyas Endale","Menberework Chanyalew"],"journal":"Scientific Reports","publisher":"Nature","impact_factor":null,"abstract":"","source_metadata":{"short_name":"Sci. Rep.","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"journals:59608bb4b4a82dfbd0263282efbdb2a23cea196e","kind":"journals","source":"Entomologia Experimentalis et Applicata","title":"General Odorant‐Binding Protein 1 Mediates the Synergistic Attractiveness of Sex Pheromones and Host Volatiles to Male in the Moth\n Phauda flammans","url":"https://doi.org/10.1111/eea.70223","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1111%2Feea.70223","date":"2026-11-14T00:00:00Z","timestamp":1794614400,"categories":["Docking & Screening"],"topic_ids":["docking"],"keywords":["molecular docking"],"matched_keywords":["molecular docking","chemical","molecular","ligands","Protein"],"matched_tags":["docking","generic_chem","biology"],"doi":"10.1111/eea.70223","external_id":"59608bb4b4a82dfbd0263282efbdb2a23cea196e","pdf_url":null,"code_url":null,"code_host":null,"authors":["Kai Lin","Jing-Hui Su","Jian-Song Wei","Xiao-Yun Wang","X. Zheng"],"journal":"Entomologia Experimentalis et Applicata","publisher":null,"impact_factor":null,"abstract":"Insects depend critically on the precise perception of chemical signals for survival and reproduction, with sex pheromones and host plant volatiles playing key roles in mate choice and host location, respectively. However, the chemosensory mechanisms underlying these processes remain poorly understood in the diurnal pest moth Phauda flammans (Lepidoptera: Phaudidae). Here we showed that male P. flammans exhibit significant attraction to both female sex pheromones and volatiles of their host plant Ficus concinna , and that a mixture of these cues exerts a synergistic attractive effect. Using qRT‐PCR, fluorescence competitive binding assays, molecular docking and RNA interference, we identify PflaGOBP1 as a protein that is highly expressed in male antennae and markedly upregulated by the pheromone component ( Z , Z , Z )‐9,12,15‐octadecatrienal and the host volatile β‐caryophyllene. This protein forms stable complexes with both ligands through static quenching, with binding primarily mediated by hydrophobic interactions. Knockdown of PflaGOBP1 by RNA interference completely abolishes male attraction to β‐caryophyllene and the mixture, while leaving the response to the sex pheromone partially intact, indicating that PflaGOBP1 is indispensable for host volatile recognition and its synergistic interaction with sex pheromones. Together, our findings reveal a molecular mechanism underlying PflaGOBP1‐mediated synergistic recognition of chemical signals and identify this protein as a potential target for behavior‐based green pest management strategies.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:6587b7c8b9d14f580fee69fd04a9ec629ca04f6c","kind":"journals","source":"International Journal of Molecular Sciences","title":"Genome-Wide CRISPR Screening Identifies NAPRT as a Key Regulator of Human Adipogenesis","url":"https://doi.org/10.3390/ijms272210216","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.3390%2Fijms272210216","date":"2026-11-14T00:00:00Z","timestamp":1794614400,"categories":[],"topic_ids":[],"keywords":["virtual screening"],"matched_keywords":["virtual screening","deep learning"],"matched_tags":["virtual_screening","machine_learning"],"doi":"10.3390/ijms272210216","external_id":"6587b7c8b9d14f580fee69fd04a9ec629ca04f6c","pdf_url":null,"code_url":null,"code_host":null,"authors":["Rui-Pu Liang","Shuang-Shuang Fan","Shu-Wen Ma","Wen-Qing Ma","Zhuo Zuo","Chun-Ge Zhong","T. Fei"],"journal":"International Journal of Molecular Sciences","publisher":null,"impact_factor":null,"abstract":"Adipocyte differentiation is central to adipose-tissue homeostasis, but its regulatory landscape remains incompletely defined in human cells. Here, we perform a genome-wide CRISPR/Cas9 knockout screen in oleic acid-differentiated human SW872 preadipocytes to identify functional regulators of white adipogenesis. The screens nominate hundreds of adipogenesis-regulating genes, and integration with human genetic evidence prioritizes nicotinate phosphoribosyltransferase (NAPRT) as a prominent player. NAPRT loss increases lipid accumulation and PPARγ abundance, reduces intracellular nicotinamide adenine dinucleotide (NAD+), and diminishes cytosolic β-catenin. Supplementation with nicotinamide mononucleotide restored NAD+ and partially reversed the signaling and adipogenic phenotypes, linking NAPRT-dependent NAD+ biosynthesis to WNT/β-catenin activity. Finally, deep learning-guided virtual screening nominates quinacrine as a candidate NAPRT inhibitor, and a lysate-based thermal shift assay shows quinacrine-dependent destabilization of NAPRT. These findings identify NAPRT as a metabolic regulator of human adipogenesis and connect NAD+ availability to WNT-dependent adipocyte fate control.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:a43fcb067a068112b7e9dbcd8303ead9e72d18d4","kind":"journals","source":"Journal of biomolecular structure & dynamics","title":"Integrated experimental and computational analysis of the binding of dihydropyrimidinone derivatives to human serum albumin","url":"https://doi.org/10.1080/07391102.2026.2756495","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1080%2F07391102.2026.2756495","date":"2026-11-14T00:00:00Z","timestamp":1794614400,"categories":["Property Prediction","Docking & Screening","ADMET & Safety","Free Energy & MD"],"topic_ids":["property","docking","admet","freeenergy"],"keywords":["molecular docking","AutoDock","QSAR","pharmacokinetic","molecular descriptors","molecular dynamics","MD simulations"],"matched_keywords":["molecular docking","AutoDock","quantitative structure-activity relationship","QSAR","pharmacokinetic","molecular descriptors","molecular dynamics","MD simulations","molecular","compounds","ligand","protein"],"matched_tags":["docking","docking_software","qsar_qspr","admet","molecular_descriptor","molecular_dynamics","generic_chem","biology"],"doi":"10.1080/07391102.2026.2756495","external_id":"a43fcb067a068112b7e9dbcd8303ead9e72d18d4","pdf_url":null,"code_url":null,"code_host":null,"authors":["Maryam Payehghadr","D. Ajloo","S. A. Pourmousavi","A. Saboury"],"journal":"Journal of biomolecular structure & dynamics","publisher":null,"impact_factor":null,"abstract":"The interaction of a series of dihydropyrimidinone (DHP) derivatives with human serum albumin (HSA) was investigated using an integrated experimental and computational approach combining fluorescence spectroscopy, quantitative structure-activity relationship (QSAR) analysis, molecular docking and molecular dynamics (MD) simulation. Fluorescence measurements performed at 27 and 37 °C under pH 4.0 and 7.0 conditions demonstrated spontaneous binding of all compounds to HSA. Binding constants and Gibbs free energies obtained from fluorescence quenching indicated that hydrophobic interactions are the principal driving force for complex formation. To establish structure-binding relationships, approximately 1,500 molecular descriptors were calculated using Dragon software. Multiple linear regression (MLR) and principal component analysis (PCA) identified molecular size, hydrophobicity, aromaticity and molecular geometry as the primary determinants of albumin binding. The QSAR models exhibited reasonable predictive performance (R = 0.93) between predicted and calculated binding free energies. Molecular docking using AutoDock 4.2 identified two major binding sites on HSA corresponding to the known ligand-binding regions and the predicted docking energies showed good agreement with experimental binding free energies (R = 0.79). MD simulations of representative high- and low-affinity DHP derivatives confirmed the stability of the protein-ligand complexes and revealed that the strongest binders formed more persistent hydrogen bonds and maintained closer interactions with HSA throughout the simulation. The close agreement among fluorescence spectroscopy, QSAR modeling, molecular docking and MD simulations provides a comprehensive molecular understanding of DHP-HSA interactions and offers valuable insights for the rational design of dihydropyrimidinone derivatives with improved plasma protein binding and pharmacokinetic properties.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:3be4eef3c5682cf227fc4701fbc7236fd28543c1","kind":"journals","source":"Atlantic Journal of Life Sciences","title":"Integrated in vitro, in vivo, and in silico evaluation of antidiabetic and anti-inflammatory activities of Melia azedarach stem bark extract","url":"https://doi.org/10.71005/e1prqf61","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.71005%2Fe1prqf61","date":"2026-11-14T00:00:00Z","timestamp":1794614400,"categories":[],"topic_ids":[],"keywords":["Molecular docking"],"matched_keywords":["Molecular docking","Molecular","protein"],"matched_tags":["docking","generic_chem","biology"],"doi":"10.71005/e1prqf61","external_id":"3be4eef3c5682cf227fc4701fbc7236fd28543c1","pdf_url":null,"code_url":null,"code_host":null,"authors":["Adil Sher","Muhammad Shahab","Shahab Ud Din","Irfan Ullah","A. Shan","Adnan Gul","Bushra Haider","M. Tauqeer","G. Rehman"],"journal":"Atlantic Journal of Life Sciences","publisher":null,"impact_factor":null,"abstract":"This study investigated the anti-inflammatory and anti-diabetic potential of Melia azedarach stem bark extract (MASBE) using integrated in vitro, in vivo, and in silico approaches. Anti-diabetic activity was assessed through yeast glucose uptake, glucose adsorption, α-amylase, and α-glucosidase inhibition assays, followed by evaluation in an alloxan-induced diabetic mouse model. Anti-inflammatory activity was evaluated by human red blood cell (HRBC) membrane stabilization, heat-induced hemolysis, protein denaturation assays, and an in vivo xylene-induced ear edema model in mice. Molecular docking investigated the interactions of major phytoconstituents with inflammatory targets (COX-1, COX-2, and 5-LOX) and the anti-diabetic target AMPK. The extract showed potent anti-diabetic effects, including maximum yeast glucose uptake (78.99%), α-amylase inhibition (76.72%), and α-glucosidase inhibition (74.53%), along with effective glucose adsorption. In diabetic mice, MASBE significantly reduced blood glucose levels and improved body weight and pancreatic and liver histology. MASBE exhibited significant dose-dependent anti-inflammatory activity, achieving maximum inhibition of 63.76% in HRBC membrane stabilization, 56.09% in heat-induced hemolysis, and 69.69% in protein denaturation assays, while markedly reducing xylene-induced ear edema in mice. Molecular docking revealed favorable binding affinities of α-linolenic acid, protocatechuic acid, and cyanidin-3-O-β-D-glucoside toward inflammatory and diabetic targets, with docking scores ranging from −3.72 to −6.85 kcal/mol, supported by key hydrogen-bond interactions. The combined experimental and computational findings demonstrate that MASBE possesses potent anti-inflammatory and anti-diabetic activities. These effects are likely mediated through multiple mechanisms involving inhibition of inflammatory mediators and modulation of glucose metabolism, highlighting M. azedarach stem bark as a promising natural source for the development of multifunctional therapeutic agents.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:0ac8de9244079784c2ade104fb65be6e5cca58fb","kind":"journals","source":"Cancer Biome and Targeted Therapy","title":"Integrative Network Toxicology and Multi-Omics Analysis Identifies Candidate Molecular Pathways Linking Acrylonitrile Exposure to Lung Cancer","url":"https://doi.org/10.66505/cbtt.v1i4.58","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.66505%2Fcbtt.v1i4.58","date":"2026-11-14T00:00:00Z","timestamp":1794614400,"categories":["Docking & Screening"],"topic_ids":["docking"],"keywords":["molecular docking","regression model"],"matched_keywords":["molecular docking","machine learning","regression model","Molecular","chemical"],"matched_tags":["docking","machine_learning","ml_task","generic_chem"],"doi":"10.66505/cbtt.v1i4.58","external_id":"0ac8de9244079784c2ade104fb65be6e5cca58fb","pdf_url":null,"code_url":null,"code_host":null,"authors":["Wen-Jing Li","Ya-Qi Hu","Chao-Fan Zhang","Tian-Song Cao","Louis Zizhao Wang","Yan-Qing Liu","Hong-Zhi Yu","Yang Cao","Jin-Rui Dong"],"journal":"Cancer Biome and Targeted Therapy","publisher":null,"impact_factor":null,"abstract":"Background: Acrylonitrile (AN) is an industrial chemical and environmental pollutant with recognized carcinogenic potential. However, the molecular pathways potentially linking AN exposure to lung cancer remain incompletely characterized. Methods: We integrated network toxicology, machine learning, SHapley Additive exPlanations (SHAP), molecular docking and dynamics simulations, and transcriptomic–metabolomic profiling to identify candidate pathways associated with AN exposure and lung cancer. Results: The intersection of predicted AN-associated targets, lung cancer-related genes, and tumor-associated differentially expressed genes identified 103 candidate genes. Enrichment analyses highlighted inflammatory and chemokine signaling, NF-κB, MAPK, and calcium signaling. Complementary machine-learning approaches prioritized EZH2, PPARG, PARP1, CXCR2, and EGFR. For tumor-versus-normal classification, the five-gene logistic regression model yielded descriptive areas under the receiver operating characteristic curve of 0.925 in the training dataset and 0.846 in the evaluation dataset. Molecular simulations generated hypotheses concerning potential AN interactions with EZH2 and EGFR. Separately, transcriptomic and metabolomic profiling of AN-exposed A549 cells identified alterations in xenobiotic metabolism, inflammatory signaling, and carbohydrate and lipid metabolism. Conclusions: These complementary analyses identify candidate molecular pathways at the intersection of AN toxicology and lung cancer biology. Because the computational analyses primarily used existing lung cancer datasets and experimental profiling was restricted to a single cancer cell line, the findings do not establish AN-specific causal mechanisms or clinically validated biomarkers. Exposure-controlled studies in additional pulmonary models and human populations are needed to evaluate the proposed pathways and their translational relevance.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:25ee2b5dde000ab36ea6881973b1319b2b486bdf","kind":"journals","source":"Chemico-biological interactions","title":"Mapping the Molecular Interaction Profile of 5-(1,3,4-Oxadiazol-2-yl)-Substituted Pyrimidine Hybrids in Diabetes Mellitus: Synthetic Development, Enzyme Kinetics, Molecular Recognition, In Vivo Toxicology and Therapeutic Translation","url":"https://doi.org/10.1016/j.cbi.2026.112428","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1016%2Fj.cbi.2026.112428","date":"2026-11-14T00:00:00Z","timestamp":1794614400,"categories":["Docking & Screening","ADMET & Safety"],"topic_ids":["docking","admet"],"keywords":["Molecular docking","Density functional theory","DFT","ADMET","pharmacokinetic","IC50","Enzyme"],"matched_keywords":["Molecular docking","Density functional theory","DFT","ADMET","pharmacokinetic","IC50","Molecular","compounds","compound","drug","chemical","Enzyme"],"matched_tags":["docking","dft","admet","bioactivity","generic_chem","biology"],"doi":"10.1016/j.cbi.2026.112428","external_id":"25ee2b5dde000ab36ea6881973b1319b2b486bdf","pdf_url":null,"code_url":null,"code_host":null,"authors":["M. M. Khowdiary"],"journal":"Chemico-biological interactions","publisher":null,"impact_factor":null,"abstract":"Diabetes mellitus is a major metabolic disorder characterized by persistent hyperglycemia and remains a significant therapeutic challenge. Inhibition of α-amylase and α-glucosidase, key enzymes involved in carbohydrate digestion, represents an effective strategy for controlling postprandial glucose levels. In the present study, a novel series of pyrimidine-oxadiazole hybrid derivatives was designed, synthesized, structurally characterized, and evaluated using complementary experimental and computational approaches. The synthesized compounds exhibited promising inhibitory activity against both enzymes, with compound 5 having IC50 = 5.20 ± 0.10 μM and 5.90 ± 0.30 μM for α-amylase and α-glucosidase, respectively,emerging as the most potent derivative against α-amylase and α-glucosidase, showing greater activity than the reference drug acarbose. Structure-activity relationship analysis revealed that the electronic and steric characteristics of substituents substantially influenced enzyme inhibitory activity. Molecular docking demonstrated favorable binding orientations. Density functional theory (DFT) calculations provided insights into the electronic properties, frontier molecular orbitals, and chemical reactivity of the active derivatives. Furthermore, in silico ADMET profiling indicated favorable drug-like and pharmacokinetic characteristics. Overall, these findings identify compound 5 and the pyrimidine-oxadiazole scaffold as promising frameworks for further development of dual α-amylase/α-glucosidase inhibitors.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:10.1021/acs.jctc.6c01711","kind":"journals","source":"Journal of Chemical Theory and Computation","title":"MemBack: An Equivariant\nGraph Neural Network for Backmapping\nLipid Membranes","url":"https://doi.org/10.1021/acs.jctc.6c01711","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1021%2Facs.jctc.6c01711","date":"2026-11-14T00:00:00+00:00","timestamp":1794614400,"categories":[],"topic_ids":[],"keywords":["Graph Neural Network","GNN","Equivariant"],"matched_keywords":["Neural Network","Graph Neural Network","GNN","Equivariant","molecular"],"matched_tags":["neural_network","gnn","equivariance","generic_chem"],"doi":"10.1021/acs.jctc.6c01711","external_id":null,"pdf_url":null,"code_url":null,"code_host":null,"authors":["Y. Eren Tunç","Rainer A. Böckmann"],"journal":"Journal of Chemical Theory and Computation","publisher":"ACS","impact_factor":null,"abstract":"Backmapping coarse-grained (CG) simulations to atomistic resolution is central to multiscale molecular simulation but remains challenging for chemically complex lipid membranes. We introduce MemBack, an SE(3)-equivariant graph neural network (GNN) that reconstructs CHARMM36 lipid structures from Martini 3 configurations by single-pass heavy-atom prediction followed by automated postprocessing. Across chemically diverse systems, MemBack achieved a mean superposition-free per-lipid heavy-atom RMSD of 0.65 Å while retaining comparable accuracy for membrane systems and lipid species excluded from training. At the membrane scale, MemBack preserved structural organization across resolutions: in a 16-component red-blood-cell (RBC) membrane model excluded from training, bilayer thickness, area per lipid (APL), and acyl-chain order closely matched the atomistic reference after only restrained minimization; in a native phase-separated Martini 3 membrane, the lateral organization of ordered and disordered domains was likewise retained directly after backmapping, without atomistic equilibration. Native Martini 3 systems containing up to 1.4 million reconstructed atoms remained consistent with their parent coarse-grained configurations and could be propagated in atomistic simulations, providing an efficient interface between Martini 3 and CHARMM36 membrane simulations.","source_metadata":{"short_name":"JCTC","source":"crossref","abstract_source":null,"abstract_quality":null}},{"id":"journals:e9f1cdeb766187e14bf46715934bc4c0df03deff","kind":"journals","source":"International Journal of Mathematics Trends and Technology","title":"Multiplicative Sombor and Euclidean Reciprocal Index and Coindex of Certain Networks","url":"https://doi.org/10.14445/22315373/ijmtt-v72i10p101","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.14445%2F22315373%2Fijmtt-v72i10p101","date":"2026-11-14T00:00:00Z","timestamp":1794614400,"categories":["Property Prediction"],"topic_ids":["property"],"keywords":["QSAR","QSPR"],"matched_keywords":["Quantitative Structure-Activity Relationship","QSAR","quantitative structure-property relationship","QSPR","molecular","molecules","molecule"],"matched_tags":["qsar_qspr","generic_chem"],"doi":"10.14445/22315373/ijmtt-v72i10p101","external_id":"e9f1cdeb766187e14bf46715934bc4c0df03deff","pdf_url":null,"code_url":null,"code_host":null,"authors":["A. S.","S. S. Sandhya"],"journal":"International Journal of Mathematics Trends and Technology","publisher":null,"impact_factor":null,"abstract":"Topological indices are mathematical numbers obtained from the molecular graph structure, which characterize properties of the molecules, such as their branched, symmetric, and connected structure. Topological indices have considerable importance in Quantitative Structure-Activity Relationship (QSAR) and quantitative structure-property relationship (QSPR) studies, which allow the prediction of the properties of the molecule without performing laborious experiments. In this paper, we introduce a new topological index called the Multiplicative Euclidean Reciprocal Index (ERI), which is the square root of the sum of reciprocals of squared degrees of adjacent nodes for all edges in the molecular graph.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:10.1021/acs.jcim.6c01918","kind":"journals","source":"Journal of Chemical Information and Modeling","title":"Non-B DNA\nFinder: A Unified Computational Framework\nfor High-Throughput Detection and Contextual Analysis of Noncanonical\nDNA Structures in Diverse Genomes","url":"https://doi.org/10.1021/acs.jcim.6c01918","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1021%2Facs.jcim.6c01918","date":"2026-11-14T00:00:00+00:00","timestamp":1794614400,"categories":["Cheminformatics"],"topic_ids":["cheminformatics"],"keywords":[],"matched_keywords":[],"matched_tags":[],"doi":"10.1021/acs.jcim.6c01918","external_id":null,"pdf_url":null,"code_url":"https://github.com/VRYella/NonBDNAFinder","code_host":"GitHub","authors":["Venkata Rajesh Yella","Aruna Sesha Chandrika Gummadi","Aditya Kumar"],"journal":"Journal of Chemical Information and Modeling","publisher":"ACS","impact_factor":5.3,"abstract":"Noncanonical DNA secondary structures arise from unique sequence patterns, and their formation in vivo is modulated by local genomic context, including DNA supercoiling, chromatin state, and transcriptional activity. They contribute fundamentally to DNA replication, transcription dynamics, genome organization, gene regulation, genetic diversity, genome instability, disease pathogenesis, and aging. However, current computational tools for non-B DNA detection are limited in scope. Most predict only one or a few structural classes. None of the tools annotate the novel, structurally distinct and biologically significant motifs characterized in recent years, viz., A-philic DNA, extended-guanine Z-DNA (eGZ-DNA), adenine-cytosine motifs (AC-motifs), and G-triplex intermediates. Furthermore, no existing tool identifies genomic regions where multiple non-B DNA conformations overlap as hybrid domains or coclusters as structural hotspots. In this study we present the Non-B DNA Finder, a high-throughput computational framework that addresses these limitations. It is the first unified platform to integrate the detection of 11 structural classes and 24 subclasses of non-B DNA motifs, encompassing both the classical non-B DNA repertoire and novel motif types. For the first time, it enables genome-wide annotation of hybrid overlap regions and non-B DNA structural clusters. To this end, we applied the pipeline to 11 complete genomes spanning a broad GC range of 17.6–76.2%. A total of ∼15 million non-B DNA motifs were identified, corresponding to an average density of ∼6 motifs/kb (range: 0.76–15.5 motifs/kb). Overall motif density was positively correlated with GC content (r = 0.37). G-quadruplexes were the most prevalent structural class in GC-rich actinobacterial genomes, reaching 7604 motifs/Mb in Miltoncostaea marina. Contrastingly, curved DNA dominated AT-rich endosymbionts, accounting for >70% of all structures in Candidatus Carsonella ruddii and Buchnera aphidicola. Importantly, hybrid interaction analysis revealed that genomic regions rarely encode isolated non-B DNA conformations. In the human genome, G-Quadruplex → i-Motif (9.57%), G-Quadruplex → Slipped DNA (9.46%), and G-Quadruplex → R-loop (5.99%) were the most preponderant hybrid pairs. Structural cluster analysis identified loci harboring up to seven co-occurring structural classes within a 300 bp window. Non-B DNA Finder is freely accessible at https://nonbdnafinder.streamlit.app/, and the source code is available at https://github.com/VRYella/NonBDNAFinder/.","source_metadata":{"short_name":"JCIM","source":"crossref","abstract_source":null,"abstract_quality":null,"code_url":"https://github.com/VRYella/NonBDNAFinder","code_status":"found"}},{"id":"journals:498ccab3255a3f316c38ed35cc438949e2b502ea","kind":"journals","source":"Journal of Advances in Medicine and Medical Research","title":"Phytochemical Characterization of Aqueous Garlic (Allium sativum) Extract by FTIR, GC–MS and Assessment of Its Antiglycation, Alpha-amylase and Antidiabetic Activities in Alloxan-induced Diabetic Sprague–Dawley Rats and Molecular Docking Study","url":"https://doi.org/10.9734/jammr/2026/v38i116247","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.9734%2Fjammr%2F2026%2Fv38i116247","date":"2026-11-14T00:00:00Z","timestamp":1794614400,"categories":["Docking & Screening"],"topic_ids":["docking"],"keywords":["Molecular Docking","binding affinity"],"matched_keywords":["Molecular Docking","Molecular","compounds","ligand","binding affinity"],"matched_tags":["docking","generic_chem","bioactivity"],"doi":"10.9734/jammr/2026/v38i116247","external_id":"498ccab3255a3f316c38ed35cc438949e2b502ea","pdf_url":null,"code_url":null,"code_host":null,"authors":["J. Momoh","M. Ekum","Mark Imhonikhe Ogbiti","Nkechi Perpetua Nwafor","O. A. Damazio","T. Oshin","O. Ishola","Cecilia Oyinlade Ogundare","Murisiku Adegbemi Onigemo","O. Adekunle","D. S. Olojede","A. Babalola","B. Kayode"],"journal":"Journal of Advances in Medicine and Medical Research","publisher":null,"impact_factor":null,"abstract":"Garlic (Allium sativum L.) contains phytochemicals of potential relevance to oxidative stress reduction and glucose homeostasis. This study characterised aqueous Allium sativum extract and evaluated its antioxidant, antiglycation and antidiabetic effects in alloxan-induced diabetic Sprague–Dawley rats, together with molecular docking against human alpha-amylase. Mineral composition, FTIR, GC–MS and phytochemical profiles were determined, while antioxidant and antiglycation activities were assessed using established in-vitro assays. Twenty-five rats were allocated to five groups. Diabetic animals received no treatment, glibenclamide (10 mg/kg), or aqueous garlic extract at 250 or 500 mg/kg body weight for 15 days, while a non-diabetic group served as the normal control. GC–MS identified 22 compounds, with 4H-pyran-4-one,2,3-dihydro-3,5-dihydroxy-6-methyl accounting for 22.53% and diallyl disulfide for 17.05%. The extract showed DPPH radical-scavenging activity with an IC₅₀ of 66.67 ± 2.07 µg/mL and inhibited advanced glycation end-product formation with an IC₅₀ of 66.75 ± 0.76 µg/mL. Treatment with the extract was associated with reduced blood glucose, total cholesterol, triglycerides, LDL-cholesterol, atherogenic indices and malondialdehyde, together with increases in HDL-cholesterol and antioxidant parameters compared with untreated diabetic animals. Pancreatic histology also showed better-preserved islet architecture in extract-treated groups. Molecular docking against human alpha-amylase (PDB: 1U2Y) yielded a binding affinity of −5.9 kcal/mol for the highest-ranked garlic-derived ligand (Hydrazine,1,2-dimethyl-), compared with −6.3 kcal/mol for acarbose. Overall, the findings indicate antioxidant, antiglycation and glucose-lowering activity of aqueous garlic extract in this experimental model.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:92de1c054fb8624bfb367f8ea10edda2c266a644","kind":"journals","source":"MASALIQ","title":"Potensi Senyawa Bioaktif dari Manilkara zapota sebagai Agen Antiporosis melalui Pendekatan In Silico terhadap Protein RANKL (3URF)","url":"https://doi.org/10.58578/masaliq.v6i6.12640","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.58578%2Fmasaliq.v6i6.12640","date":"2026-11-14T00:00:00Z","timestamp":1794614400,"categories":["Docking & Screening","ADMET & Safety"],"topic_ids":["docking","admet"],"keywords":["ADMET prediction","drug likeness","molecular docking","Receptor","ADMET","pharmacokinetic","binding affinity"],"matched_keywords":["ADMET prediction","drug-likeness","molecular docking","Protein","Receptor","ADMET","pharmacokinetic","ligand","compounds","molecular","ligands","drug","binding affinity"],"matched_tags":["admet_prediction","drug_likeness","docking","biology","admet","generic_chem","bioactivity"],"doi":"10.58578/masaliq.v6i6.12640","external_id":"92de1c054fb8624bfb367f8ea10edda2c266a644","pdf_url":null,"code_url":null,"code_host":null,"authors":["Intan Elwis Khairina Muzia","A. Ulianas"],"journal":"MASALIQ","publisher":null,"impact_factor":null,"abstract":"Osteoporosis is characterized by reduced bone mass and an imbalance between bone formation and resorption. Receptor activator of nuclear factor kappa-B ligand (RANKL) plays an important role in osteoclast differentiation, making it a potential target for the development of osteoporosis therapies. This study aimed to evaluate the potential of bioactive compounds from Manilkara zapota as candidate antiosteoporosis agents through an in silico approach using molecular docking against the RANKL protein (PDB ID: 3URF). Ten bioactive compounds were used as test ligands, with denosumab as the positive control. The study stages included protein and ligand preparation, redocking validation, molecular docking, drug-likeness evaluation, and ADMET prediction. Redocking yielded an RMSD value of 0.8827 Å. All test compounds yielded negative binding affinity scores against RANKL, indicating potential ligand–protein interactions based on computational modeling. Most compounds met the drug-likeness criteria and were predicted to have favorable ADMET profiles, whereas myricetin-3-O-glucoside and chlorogenic acid did not meet several drug-likeness parameters. These findings provide a preliminary computational basis for exploring bioactive compounds from M. zapota as candidate antiosteoporosis agents. However, docking scores and predicted pharmacokinetic properties do not yet demonstrate the compounds’ biological activity, effectiveness, or safety, making further in vitro and in vivo testing necessary.","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:1f565390baa389c0a112d02876394662273b4d6f","kind":"journals","source":"Spectroscopy Letters","title":"Structural, spectroscopic (FT-IR, FT-Raman, UV-vis, NMR), DFT, molecular docking and\n in vitro\n analysis on 2-(4-chlorobenzoyl)benzoic acid: liver cancer inhibitor","url":"https://doi.org/10.1080/00387010.2026.2761861","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.1080%2F00387010.2026.2761861","date":"2026-11-14T00:00:00Z","timestamp":1794614400,"categories":["Docking & Screening"],"topic_ids":["docking"],"keywords":["molecular docking","DFT"],"matched_keywords":["molecular docking","DFT","molecular"],"matched_tags":["docking","dft","generic_chem"],"doi":"10.1080/00387010.2026.2761861","external_id":"1f565390baa389c0a112d02876394662273b4d6f","pdf_url":null,"code_url":null,"code_host":null,"authors":["S. Vanitha","S. Sebastian","S. Sylvestre","S. Silvan"],"journal":"Spectroscopy Letters","publisher":null,"impact_factor":null,"abstract":"","source_metadata":{"source":"semantic_scholar"}},{"id":"journals:7bd48f9291df8fe8ebd4da1decbd7a6aec5769dd","kind":"journals","source":"Jurnal Medika Nusantara","title":"Studi in Silico Senyawa Ekstrak Etanol Bunga Mentega (Nerium Oleander L.) terhadap Reseptor Estrogen Alfa dan Cox-2 sebagai Target Molekuler yang Berkaitan dengan Kanker Payudara dan Inflamasi","url":"https://doi.org/10.59680/medika.v4i4.2895","detail_url":"/chemradar/article?u=https%3A%2F%2Fdoi.org%2F10.59680%2Fmedika.v4i4.2895","date":"2026-11-14T00:00:00Z","timestamp":1794614400,"categories":["Docking & Screening"],"topic_ids":["docking"],"keywords":["Lipinski","molecular docking","AutoDock","molecular dynamics","receptor"],"matched_keywords":["Lipinski","molecular docking","AutoDock","compounds","molecular","molecular dynamics","receptor"],"matched_tags":["drug_likeness","docking","docking_software","generic_chem","molecular_dynamics","biology"],"doi":"10.59680/medika.v4i4.2895","external_id":"7bd48f9291df8fe8ebd4da1decbd7a6aec5769dd","pdf_url":null,"code_url":null,"code_host":null,"authors":["N. Zein","T. Tukiran"],"journal":"Jurnal Medika Nusantara","publisher":null,"impact_factor":null,"abstract":"Nerium oleander L. contains a variety of secondary metabolites, yet the potential interactions of these compounds with molecular targets related to breast cancer and inflammation remain poorly characterized. This study aimed to investigate, in silico, the compounds found in the ethanol extract of Nerium oleander flowers, focusing on their interaction with estrogen receptor-alpha (ER-a PDB ID: 3ERT)-a molecular target associated with breast cancer-and cyclooxygenase-2 (COX-2; PDB ID: 5IKT)-a molecular target involved in inflammatory pathways. Through LC-MS analysis, 34 dominant compounds with a relative abundance of at least 1% were tentatively identified and subsequently screened using Lipinski's Rule of Five. Nine compounds were then selected for molecular docking using AutoDock 4 (via AutoDockTools 1.5.7). Redocking validation yielded an RMSD of 1.232 A for 3ERT and 1.407 A for 5IKT. Regarding These values are computational predictions only and do not confirm anticancer or anti-inflammatory activity; the compounds are therefore prioritized for further validation through molecular dynamics, in vitro assays, and toxicity testing.","source_metadata":{"source":"semantic_scholar"}},{"id":"preprints:2611.18490v1","kind":"preprints","source":"arXiv","title":"Fine-tuned atomistic foundation model for accurate sodium-carbon chemistry in Na-ion batteries","url":"https://arxiv.org/abs/2611.18490v1","detail_url":"/chemradar/article?u=https%3A%2F%2Farxiv.org%2Fabs%2F2611.18490v1","date":"2026-11-13T18:04:13Z","timestamp":1794593053,"categories":["Property Prediction","ML Potentials"],"topic_ids":["property","potentials"],"keywords":["MACE","DFT"],"matched_keywords":["MACE","DFT","chemical"],"matched_tags":["mlip_tools","dft","generic_chem"],"doi":null,"external_id":"2611.18490v1","pdf_url":"https://arxiv.org/pdf/2611.18490v1","code_url":null,"code_host":null,"authors":["Andres Felipe Usuga","Ihor Radchenko","Yuxin Tang","Oleksandr I. Malyi"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"Atomistic foundation models have emerged as a promising route to DFT-level accuracy for materials simulation with complex chemical environments. However, their reliability across competing bonding regimes and, more importantly, how complex bonding affects predictions of material properties remain unclear. Among materials with competing bonding regimes, sodium storage in hard carbon poses a substantial challenge. Hard carbon features diverse local structural motifs, including defective and curved graphitic regions, amorphous regions, and closed/open pores. Motivated by its importance for Na-ion batteries, we systematically evaluate existing pretrained foundation models based on the MACE architecture across different local Na/C motifs. We find that the pretrained models fail to accurately capture Na-Na and Na-C interactions for metallic Na clusters on graphene. To recover a DFT-level energetic description, we employ targeted fine-tuning, showing that it can describe the energetic stability of local motifs and achieve sufficient transferability to unseen Na/C environments. With the developed fine-tuned model, we establish a correlation between closed-pore radius in hard carbon and plateau voltage, showing that the small pores contribute at larger voltages than large ones, suggesting thus how the developed fine-tuned foundation model can link atomistic information with materials design for complex bonding environments.","source_metadata":{"categories":["cond-mat.mtrl-sci","physics.comp-ph"],"code_status":"absent"}},{"id":"preprints:2611.18236v1","kind":"preprints","source":"arXiv","title":"Deciphering multiple binding modes in large protein complexes via Enhanced Sampling and beta-VAE: the CYFIP1-eIF4E case","url":"https://arxiv.org/abs/2611.18236v1","detail_url":"/chemradar/article?u=https%3A%2F%2Farxiv.org%2Fabs%2F2611.18236v1","date":"2026-11-13T14:31:47Z","timestamp":1794580307,"categories":["Design de novo","Free Energy & MD"],"topic_ids":["denovo","freeenergy"],"keywords":["VAE","Variational Autoencoder","Molecular Dynamics"],"matched_keywords":["VAE","Variational Autoencoder","Molecular","Molecular Dynamics","protein"],"matched_tags":["vae","generic_chem","molecular_dynamics","biology"],"doi":null,"external_id":"2611.18236v1","pdf_url":"https://arxiv.org/pdf/2611.18236v1","code_url":null,"code_host":null,"authors":["Francesco Pignotti","Emanuele Telari","Edoardo Milanetti","Anna La Teana","Jesmina Rexha","Alice Romagnoli","Luca Maragliano","Giovanni Di Muccio"],"journal":null,"publisher":null,"impact_factor":null,"abstract":"Eukaryotic initiation factor 4E (eIF4E) drives cap-dependent translation, regulated by 4E-binding proteins (4E-BPs) competing for its dorsal and lateral surfaces. CYFIP1, a scaffold protein, is both a WAVE Regulatory Complex component and a non-canonical 4E-BP repressing translation via a short helical motif, linked to autism spectrum disorders and cancer. Prior work modeled only this isolated motif as a peptide, leaving the full interface uncharacterized. Starting from this preliminary evidence, we modeled the full-length proteins as flexible partners: all-atom simulations fed a data-driven docking search (HADDOCK3), which guided funnel-restrained Temperature-Accelerated Molecular Dynamics of a coarse-grained Martini 3 representation. Ten replicas accumulated 100 $μ$s of enhanced sampling. Funnel-restraint variables efficiently sample binding and unbinding but are structurally degenerate: direct clustering of the raw trajectory collapses these poses into poorly resolved groups. We instead reweighted the biased trajectories onto a contact-based latent space built by benchmarking PCA against a beta-Variational Autoencoder, then clustered minima of the equilibrium free-energy surface: only intermediate regularization resolved distinct, homogeneous modes. Three such modes recur, engaging eIF4E's canonical dorsal site, its non-canonical lateral site, and both. We find that CYFIP1 also engages eIF4E through a second surface, the same that mediates its WAVE-complex dimerization, and a basic patch. Both plausibly explain residual binding previously reported after motif mutation. This suggests its translational and cytoskeletal roles compete for the same surface. Most contacts on eIF4E fall outside this dorsal/lateral interface. This heterogeneous, plastic ensemble offers a structural basis for mutagenesis and therapeutic peptide design.","source_metadata":{"categories":["q-bio.BM","q-bio.QM"],"code_status":"absent"}}],"total_article_count":3258,"filters":{"topic":"all","source":"all","days":7,"limit":null}}