Molecular Informatics最新文献

{"title":"Data-Driven Prediction of Ultraviolet-Visible Spectra From Molecular Structure Using Attention-Based Graph Neural Networks.","authors":"Nguyen Van Phuong, Ngo Huy Trung, Nguyen Trung Khoa","doi":"10.1002/minf.70056","DOIUrl":"https://doi.org/10.1002/minf.70056","url":null,"abstract":"<p><p>Predicting UV-vis spectra from molecular structure is an important but relatively underexplored task. In this study, we developed a deep learning approach based on graph neural networks (GNNs) to directly predict UV-vis spectra from SMILES representations. Several GNN architectures were evaluated, among which Attentive Fingerprint achieved the best performance with an R² of 0.8630, mean absolute error of 0.0389, root mean squared error of 0.0740, cosine similarity of 0.9589, Pearson correlation coefficient of 0.9521, and distance metric of 1.5758 on the test set. Further analysis showed that the model might be able to capture meaningful structure-property relationship across diverse chemical classes. The model was also evaluated on external datasets, where it maintained good predictive performance, indicating its generalizability. These results demonstrate that attention-based GNNs provide an effective approach for UV-vis spectrum prediction and may offer useful insights into the molecular regions contributing to model predictions, with potential applications in molecular analysis, high-throughput screening, and compound design. To facilitate practical use, we deployed the model as an interactive web application (https://spectra-prediction.streamlit.app/), which enables direct prediction of UV-vis spectra from SMILES inputs.</p>","PeriodicalId":18853,"journal":{"name":"Molecular Informatics","volume":"45 10","pages":"e70056"},"PeriodicalIF":3.1,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149081477","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Graph-Based Nitrogen Substitution and CatBoost Regression for Virtual Screening of Small-Molecule Acceptors in Organic Photovoltaics.","authors":"Bo Xiao, Jindu Li, Ahmed M Fallatah, Mohamed M Ibrahim, Amal M Alkhudaydi","doi":"10.1002/minf.70057","DOIUrl":"https://doi.org/10.1002/minf.70057","url":null,"abstract":"<p><p>Machine learning-driven high-throughput screening strategy was used for designing efficient nitrogen-containing small-molecule acceptors (SMAs) for organic solar cells. First, a database containing 1027 experimentally determined SMAs was encoded using RDKit molecular descriptors with the application of variance filter, correlation filter, and consensus feature selection. Then, 11 different regression algorithms were assessed in terms of PCE prediction; the best results were obtained by CatBoost algorithm. Afterward, a graph-based nitrogen substitution approach was implemented to design about 200 K chemically valid mono-, di-, and tri-nitrogen-containing SMA candidates. Further virtual screening allowed us to identify several promising candidates possessing predicted PCE value up to 16.42%. The structural analysis of chemical space revealed that all candidates are diverse and suitable for further investigation.</p>","PeriodicalId":18853,"journal":{"name":"Molecular Informatics","volume":"45 10","pages":"e70057"},"PeriodicalIF":3.1,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149183164","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"LigGen: A Tool for Predicting Water Positions in Proteins.","authors":"Aleksey Shustikov","doi":"10.1002/minf.70059","DOIUrl":"https://doi.org/10.1002/minf.70059","url":null,"abstract":"<p><p>To create a drug-like chemical space for a specific binding pocket of protein,it is important to take into account not only the topology of the protein's amino acids, but also their local environment-at least water molecules. In this article, we have written a Python program that predicts the position of water molecules (oxygen atoms) in protein crystals. The algorithm also sorts the predicted oxygen atoms based on their polarity coefficients. The user can change: the length of the hydrogen bonds between the most electronegative atoms (N, O, S) of amino acids and water oxygen; the accuracy of prediction; number of processor cores as well as program parameters. LigGen is available at: https://github.com/e-shstkv/LigGen.</p>","PeriodicalId":18853,"journal":{"name":"Molecular Informatics","volume":"45 10","pages":"e70059"},"PeriodicalIF":3.1,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149112373","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Identification of YES1 Inhibitors From Approved Drugs Using a Multimodal Deep Learning Framework.","authors":"Takayoshi Kawabata, Tatsuaki Takeda, Yuta Tanaka, Soichiro Ushio, Akira Miyamoto, Koki Nakagomi, Tomonori Sakai, Risa Maruyama, Takashi Makita, Hirofumi Hamano, Shinichi Toyooka, Yoshito Zamami","doi":"10.1002/minf.70061","DOIUrl":"10.1002/minf.70061","url":null,"abstract":"<p><p>While anti-human epidermal growth factor receptor 2 (HER2) drugs, including trastuzumab improve the prognosis of patients with HER2-positive cancers, acquired resistance by cancer cells remains a significant clinical challenge. We had previously reported that YES1 is overexpressed in trastuzumab-resistant breast and lung cancer cells, and that its inhibition restores trastuzumab sensitivity. However, no approved drug specifically targets YES1. In this study, as a drug repositioning approach aimed at rapid clinical application, we constructed a machine learning model capable of learning the interactions between compounds and target proteins and conducted a screening of approved drugs. The strategy identified chlorpromazine and raloxifene as hit compounds. In vitro validation demonstrated that they effectively inhibited YES1 activation at 15 μM after 48 h. The findings suggested that the identified approved drugs might represent promising therapeutic options for overcoming anti-HER2 drug resistance.</p>","PeriodicalId":18853,"journal":{"name":"Molecular Informatics","volume":"45 10","pages":"e70061"},"PeriodicalIF":3.1,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13665583/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149183193","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Property-Conditioned Transformer Language Model for De Novo Molecular Generation Using Self-Referencing Embedded Strings.","authors":"Ashish Patel, Rajesh A Maheshwari, Atul Patel, Ankit Faldu, Anjali Mahavar","doi":"10.1002/minf.70055","DOIUrl":"https://doi.org/10.1002/minf.70055","url":null,"abstract":"<p><p>Transformer-based molecular language models have demonstrated strong capabilities for de novo molecular generation by modeling chemical structures as token sequences. However, controllable generation of molecules with desired physicochemical properties remains challenging without reinforcement learning (RL) or post-generation filtering. In this study, we propose a property-conditioned autoregressive transformer framework for molecular generation using the SELFIES representation. Continuous molecular descriptors, including logP, quantitative estimate of drug-likeness, and molecular weight (MolWt), were normalized and embedded via a multilayer perceptron and injected into token embeddings to enable global conditioning during sequence modeling. The model was trained end-to-end using the maximum likelihood with cross-entropy loss. Experiments conducted on the combined subsets of QM9, ZINC, and MOSES datasets demonstrated high validity (100%), strong uniqueness, and substantial novelty in the generated molecules. Furthermore, the generated property distributions closely aligned with the training data statistics. The proposed framework provides a computationally efficient and stable approach for controllable molecular generation without requiring RL or graph-based modeling.</p>","PeriodicalId":18853,"journal":{"name":"Molecular Informatics","volume":"45 10","pages":"e70055"},"PeriodicalIF":3.1,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149075413","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"SMILES2Docking: An Open-Source Desktop Workflow for Ligand Ionization, Stereochemistry-Aware Preparation and Semi-Empirical 3D Refinement.","authors":"Daniel Andrés Grajales Ruiz, Bruna Flôres Negrisoli, Isabel Cristina Conceição Periquito, Nailton Monteiro do Nascimento-Júnior, Adriano Marques Gonçalves","doi":"10.1002/minf.70053","DOIUrl":"10.1002/minf.70053","url":null,"abstract":"<p><p>SMILES2Docking converts spreadsheet-based SMILES libraries into docking-ready 3D ligands through a single, configurable pipeline. Salt and coformer removal, a three-mode stereocenter policy, prediction of the dominant protonation state with a graph neural network pKa model (MolGpKa) refined by iterative titration, distance-geometry embedding with RDKit, a molecular-mechanics optimization cascade, and an optional semi-empirical refinement under an implicit solvation model are combined in one tool, with records processed either sequentially or distributed across CPU cores. Users select among permissive, strict, and enumerative stereocenter modes, so the same tool serves broad exploratory screening and precision docking. A per-compound JSON audit report records the ionization and stereocenter decisions taken for every structure. The tool is distributed as a Python package, a Windows executable with bundled MOPAC, a Linux portable bundle, and a macOS bundle, under GPL-2.0-or-later, and is freely available to non-commercial users at https://github.com/amgoncalvesusp/Smiles2Docking (DOI: 10.5281/zenodo.20617898).</p>","PeriodicalId":18853,"journal":{"name":"Molecular Informatics","volume":"45 9","pages":"e70053"},"PeriodicalIF":3.1,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13550955/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148906792","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"MSI: A Mahalanobis-Based Molecular Similarity Index for High-Dimensional Embeddings.","authors":"Roberto Bernal-Jaquez, Elliot Ridout-Buhl, Emiliano Montoya, León Alday-Toledo, Felipe Aparicio","doi":"10.1002/minf.70051","DOIUrl":"10.1002/minf.70051","url":null,"abstract":"<p><p>Quantifying molecular similarity is crucial for drug discovery and for exploring chemical space. A similarity assessment always combines two independent ingredients: a molecular representation and a similarity coefficient. The most common pairing, binary substructure fingerprints scored with the Tanimoto coefficient, depends strongly on fingerprint bit density and, because it compares unweighted sets of substructure identifiers, is blind to the multiplicity of repeated fragments and frequently returns ranking ties that obscure meaningful chemical relationships. Here, we introduce the Mahalanobis Similarity Index (MSI), which pairs continuous Mol2Vec embeddings with a covariance-aware Mahalanobis distance (d_M) and an associated Mahalanobis angle (θ_M) to give a statistically grounded assessment that is invariant under invertible linear reparametrization of the descriptor space. We evaluated MSI on five chemically distinct reference compounds: aspirin, a salicylate nonsteroidal anti-inflammatory drug (NSAID); aniline, an industrial aromatic amine; curcumin, a polyphenolic natural product; ibuprofen, a propionic-acid NSAID; and digitoxin, a cardiac glycoside. Relative to the Tanimoto coefficient computed on ECFP4 fingerprints, MSI improves the analysis in three specific respects: it promotes chemically reasonable analogs that the fingerprint deprioritises; it resolves ranking ties, recovering between 7 and 10 distinct scores among the 10 nearest neighbors where Tanimoto recovers only 3-7; and its geometry varies systematically with HOMO-LUMO energy gaps in the QM9 dataset, indicating that the embedding tracks electronic structure even though it was trained on structural context alone. Polar plots and three-dimensional similarity maps reveal anisotropy within the embedding space and define practical applicability domains for high-similarity retrieval. MSI retains discriminatory power in the regime where the Tanimoto coefficient saturates near zero, and sparse peripheral regions suggest scaffold-hopping opportunities. The dual radial-angular description supports hypothesis-driven reasoning about how structural modifications shift electronic properties. MSI is computationally efficient, chemically interpretable, and offers a practical way to navigate high-dimensional chemical space. Beyond drug discovery, it is applicable to materials science, toxicology, and chemical biology, wherever continuous molecular embeddings are used for property-driven screening.</p>","PeriodicalId":18853,"journal":{"name":"Molecular Informatics","volume":"45 9","pages":"e70051"},"PeriodicalIF":3.1,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13522730/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148840830","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Predicting the Toxicity of Chemical Compounds via Hyperdimensional Computing.","authors":"Fabio Cumbo, Kabir Dhillon, Jayadev Joshi, Bryan Raubenolt, Davide Chicco, Sercan Aygun, Daniel Blankenberg","doi":"10.1002/minf.70052","DOIUrl":"10.1002/minf.70052","url":null,"abstract":"<p><p>Accurately and efficiently assessing the potential toxicity of chemical compounds is critical given their wide application across pharmaceutical, industrial, and environmental domains. Traditional toxicological evaluations, which predominantly rely on intensive in vitro and in vivo assays, are frequently slow and expensive. Here, we introduce a novel application of hyperdimensional computing (HDC), a recently developed computational paradigm inspired by the way the human brain works in encoding information, for the efficient classification of chemical compounds as either toxic or nontoxic. Our methodology employs Simplified Molecular Input Line Entry System (SMILES) representations of compounds, drawing data from the comprehensive Tox21 dataset. We delineate a pipeline wherein these chemical structures are encoded into high-dimensional binary vectors, which subsequently serve as the foundation for training and classification within the HDC framework. This approach leverages HDC's inherent advantages, including its resilience to noise, parallel processing capabilities, and efficacy in identifying intricate patterns. This work demonstrates the viability of HDC as a computationally lightweight first-pass solution for preliminary toxicity screening. This research significantly contributes to the field of cheminformatics by validating HDC's potential in chemical property prediction, thereby facilitating accelerated identification of hazardous substances and mitigating the reliance on intensive laboratory experimentation.</p>","PeriodicalId":18853,"journal":{"name":"Molecular Informatics","volume":"45 9","pages":"e70052"},"PeriodicalIF":3.1,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148912871","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Structural and Biological Evidence for Acetylcholinesterase-Targeted Insecticidal Activity in Musca domestica: A 4.5  Å Distance Correlation From Novel Benzenesulfonamide Derivatives.","authors":"Mohamed A Bakry, Asmaa A Kamel, Lilian N Mahrous, Fatma I Abo El-Ela, Wael A A Fadaly, Shawky M Aboelhadid","doi":"10.1002/minf.70054","DOIUrl":"10.1002/minf.70054","url":null,"abstract":"<p><p>House fly (Musca domestica) control faces mounting resistance to conventional insecticides, driving interest in novel chemistries targeting acetylcholinesterase (AChE). We examined eight benzenesulfonamide derivatives originally developed as carbonic anhydrase inhibitors to explore their potential as AChE-targeting insecticides and to identify structural determinants of biological activity. Using a high-quality homology model of M. domestica AChE, we performed blind docking, refined docking to the peripheral anionic site, 100 ns molecular dynamics simulations, and molecular mechanics-poisson-boltzmann surface area (MM-PBSA) binding energy calculations. All eight compounds showed favorable binding energies (-8.5 to -10.2 kcal/mol) and stable trajectories, yet bioassays revealed a stark divide: five compounds (16b, 16c, 16d, 19b, and 19f) produced no mortality at 5% concentration, while 14g and 17c achieved complete adult kill at 5% with LC₅₀ values of 1.37% (95% CI: 1.15%-1.63%) and 1.64% (95% CI: 1.38%-1.95%), respectively. The critical difference was geometric, not thermodynamic. Active compounds 14g and 17c penetrated the ~20 Å aromatic gorge to position within 2.5-2.8 Å of the catalytic serine (Ser318), forming hydrogen bonds and halogen bonds with the triad. Inactive compounds arrested at 7.0-7.5 Å, trapped in the peripheral anionic site with negligible catalytic residue contact despite favorable overall binding. This 4.5 Å threshold correlated with biological activity and provided a quantitative filter for prioritizing AChE inhibitor candidates. Compound 14g, bearing 3,4-dichloro substituents, showed the closest approach (2.5 Å) and strongest binding energy (-37.4 kcal/mol), outperforming chlorpyrifos (-32.7 kcal/mol). Importantly, whole-organism mortality does not prove direct AChE inhibition; alternative mechanisms including cuticular penetration, metabolic disruption, or off-target effects cannot be excluded without direct enzymatic assays. These findings suggest that geometric distance to the catalytic triad, not binding energy alone, correlates with biological activity and provides a quantitative filter for prioritizing AChE inhibitor candidates.</p>","PeriodicalId":18853,"journal":{"name":"Molecular Informatics","volume":"45 9","pages":"e70054"},"PeriodicalIF":3.1,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13575437/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148970176","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Boosting the Prediction Accuracy of Glass Transition Temperature in Polyimides: A Hybrid Machine Learning Approach Integrating Morgan Fingerprints and Molecular Descriptors.","authors":"Peishuai Xing, Xiaodong Guo, Yang Wang, Zeting Chen, Shicheng Shen, Bin He, Tianyun Li, Dongmei Peng, Zhenhuai Yang","doi":"10.1002/minf.70049","DOIUrl":"10.1002/minf.70049","url":null,"abstract":"<p><p>The glass transition temperature (T_g) of polyimides is a critical parameter determining their processability and application performance. Traditional experimental methods for measuring T_g are time-consuming and costly, while existing machine learning prediction models predominantly rely on manually defined molecular descriptors, which often fail to fully capture detailed molecular structural information, limiting their prediction accuracy and generalization capability. To address this, this study proposes a hybrid feature engineering strategy combining Morgan fingerprints and molecular descriptors to comprehensively represent the chemical structure of polyimides. Based on a dataset of 1257 polyimide samples from a public database, we systematically compared six feature selection methods and employed multiple mainstream machine learning algorithms for modeling. The results show that the CATB model performed best, achieving a coefficient of determination (R²) of 0.882 and a mean absolute error (MAE) of 17.34 °C on an independent test set, with fivefold cross-validation further confirming the model's robustness. SHAP interpretability analysis revealed the significant influence of key features such as the number of rotatable bonds, ether bonds, and ether-linked oxyethylene units on T_g, providing clear guidance for molecular design. External validation demonstrated the model's strong generalization ability. This study not only achieves high-precision and robust T_g prediction but also highlights the importance of hybrid feature strategies in polymer property modeling, offering a data-driven foundation for the rational design of polyimides.</p>","PeriodicalId":18853,"journal":{"name":"Molecular Informatics","volume":"45 8","pages":"e70049"},"PeriodicalIF":3.1,"publicationDate":"2026-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13500066/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148808763","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}