Journal of Cheminformatics最新文献

{"title":"Transformer-accelerated discovery of inhibitors targeting the RpsAΔ438 deletion in PZA-resistant tuberculosis","authors":"Jiazhuo Xiao, Muhammad Shahab, Haoke Zhang, Zunnan Huang","doi":"10.1186/s13321-026-01235-0","DOIUrl":"10.1186/s13321-026-01235-0","url":null,"abstract":"<div><p>Tuberculosis is a major global public health problem caused by <i>Mycobacterium tuberculosis</i>. Pyrazinamide (PZA), a key first-line drug, is activated to pyrazinoic acid, which inhibits <i>Mycobacterium tuberculosis</i> by targeting ribosomal protein S1 (RpsA) and blocking trans-translation. Deletion of alanine at position 438 of RpsA (RpsA<sub>Δ438</sub>) increases C-terminal flexibility, reduces drug binding, and confers PZA resistance. To identify novel inhibitors against this mutant, we developed a screening pipeline integrating Transformer-based virtual screening, molecular docking, Gaussian accelerated molecular dynamics simulations, and MM/GBSA free energy calculations. Compounds from the ChEMBL database were curated, labeled by MIC, and used to train Transformer-based models. The best-performing Transformer model outperformed Forest and XGBoost across multiple evaluation metrics, demonstrated an AUC of 0.844 and an accuracy of 0.772 on the test set, was applied to screen FDA-approved and traditional Chinese medicine libraries, yielding 317 and 1757 candidate compounds, respectively. Top-ranked hits were docked into RpsA<sub>Δ438</sub> and evaluated using 400 ns molecular dynamics and Gaussian accelerated molecular dynamics simulations. Trajectory analyses revealed that compound Z2568748600 was the most promising candidate, exhibiting an RMSD of approximately 1.2–2.0 Å, an RMSF of 1.5–2.0 Å, and a radius of gyration of ~18.0 Å, forming long-lived hydrogen bonds with multiple residues, markedly suppressing large-scale motions, and effectively stabilizing the deletion-affected C-terminal domain. The MM/GBSA binding free energy was calculated to be −26.00 ± 3.50 kJ/mol. Moreover, in silico ADMET profiling confirmed its favorable drug-likeness and balanced safety profile, underscored by excellent cardiac safety and minimal cellular cytotoxicity. Overall, this integrated approach offers a promising strategy for discovering new agents against PZA-resistant <i>Mycobacterium tuberculosis</i>.</p><p><b>Scientific contribution</b></p><p>This study established an integrated, systematic screening strategy targeting the drug-resistant RpsA<sub>Δ438</sub> protein by coupling a Transformer-based deep learning framework for high-throughput virtual screening with subsequent molecular docking, Gaussian accelerated molecular dynamics simulations, MM/GBSA free energy calculations and ADMET prediction. This strategy enabled the efficient identification of potential inhibitors from expansive chemical space and allowed for the in-depth evaluation of compound interactions with the mutant protein. Among the screened candidates, Z2568748600 was identified as the most promising compound based on comprehensive computational assessments. Collectively, this approach significantly improved the discovery efficiency of promising inhibitors against pyrazinamide-resistant <i>Mycobacterium tuberculosis</i>.</p><h3>Graphical Abstract</h3>\u0000<div><figure><div><div><pi","PeriodicalId":617,"journal":{"name":"Journal of Cheminformatics","volume":"18 1","pages":""},"PeriodicalIF":7.9,"publicationDate":"2026-06-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://link.springer.com/content/pdf/10.1186/s13321-026-01235-0.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148299079","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"化学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Systematic validation of graph neural network explanations against adverse outcome pathway reactive centers for skin sensitization","authors":"ByoungJun Jeon","doi":"10.1186/s13321-026-01249-8","DOIUrl":"10.1186/s13321-026-01249-8","url":null,"abstract":"<p>Whether graph neural network (GNN) attributions capture the reactive chemistry encoded in adverse outcome pathway (AOP) annotations for skin sensitization has not been tested under a label-permutation control. We trained AttentiveFP on a 436-molecule LLNA-labeled subset (81 sensitizers) of a curated skin-sensitization dataset and extracted atom-level attributions from seven methods (integrated gradients, GradCAM, attention, GNNExplainer, PGExplainer, GraphMask, and a rank-aggregated ensemble); GCN and GIN were retrained as architecture controls. Attributions were evaluated against 50 sensitization-specific SMARTS-based reactive-center annotations spanning six molecular initiating event (MIE) mechanism classes. The LLNA-trained model reached test AUC 0.73, compared with AUC 0.49 for a shuffled-label retraining control. Across all seven attribution methods, paired real-vs-shuffled atom-AUC differences crossed zero, indicating that most apparent attribution–mechanism correlation reflects structural priors of the attribution methods rather than learned sensitization chemistry. Mechanism-class stratification surfaced one candidate exception—pre-hapten autoxidation chemistry, with a <span>(Delta = +0.26)</span> atom-AUC gap and non-overlapping confidence intervals—on a stratum of approximately two unique molecules, reported as a hypothesis pending replication on dedicated pre-hapten benchmarks. Per-element residual analysis converges on the same interpretation: alignment for gradient- and attention-based methods drops to chance once elemental identity is regressed out.</p><p>This work pairs a reproducible evaluation framework (50 skin-sensitization-specific SMARTS patterns scored against atom-level GNN attributions) with a label-permutation negative control that separates learned chemistry from structural priors of the attribution method.</p>","PeriodicalId":617,"journal":{"name":"Journal of Cheminformatics","volume":"18 1","pages":""},"PeriodicalIF":7.9,"publicationDate":"2026-06-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://link.springer.com/content/pdf/10.1186/s13321-026-01249-8.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148289717","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"化学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"DICL: a manually curated database of ion channels and ligands as a useful platform for drug discovery targeting ion channels","authors":"Wei Zhao,&nbsp;Xiaoqing Gong,&nbsp;Yanan Tian,&nbsp;Lingling Wang,&nbsp;Chenggong Fu,&nbsp;Xiaojun Yao,&nbsp;Huanxiang Liu","doi":"10.1186/s13321-026-01246-x","DOIUrl":"10.1186/s13321-026-01246-x","url":null,"abstract":"<div><p>Ion channels represent a crucial class of drug targets. Currently, with the structural elucidation of more and more ion channels and the rapid development of artificial intelligence in drug discovery, unprecedented opportunities have emerged for designing and discovering drugs that target ion channels. However, both traditional computer-aided drug design and artificial intelligence drug design rely heavily on a large amount of effective data, particularly high-quality data. Yet, data pertaining to relevant ion channels and their ligands are notably scarce. Therefore, constructing a comprehensive and reliable database encompassing ion channels and their ligands is paramount for drug discovery efforts centered around ion channels. To address this gap, we developed database (DICL, Database of Ion Channels and Ligands) including 464,225 data records with the manually curated structure and activity information of ligands from over 6,000 peer-reviewed publications, patents and other databases. As far as we know, it represents the largest and most detailed online database dedicated to ion channels and their ligands. Additionally, the database integrates information regarding the binding sites of ligands, channelopathies, and specific experimental conditions related to biological activities for the first time. Furthermore, it features a user-friendly interface that facilitates querying, browsing, and utilization of this resource. DICL is freely available and can be accessed at https://fca_icdb.mpu.edu.mo/ICHDB/public/. </p><h3>Graphical abstract</h3>\u0000<div><figure><div><div><picture><source><img></source></picture></div></div></figure></div></div>","PeriodicalId":617,"journal":{"name":"Journal of Cheminformatics","volume":"18 1","pages":""},"PeriodicalIF":7.9,"publicationDate":"2026-06-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://link.springer.com/content/pdf/10.1186/s13321-026-01246-x.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148289672","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"化学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"DCPM-ADMET: fusion of dual-component pre-trained model and molecular fingerprints to enhance drug ADMET properties prediction","authors":"Leilei Zhang,&nbsp;Yuchen Zeng,&nbsp;Yue Qi,&nbsp;Kaili Jiang,&nbsp;Xiaofei Zhou,&nbsp;Lu Liang,&nbsp;Jianping Lin","doi":"10.1186/s13321-026-01244-z","DOIUrl":"10.1186/s13321-026-01244-z","url":null,"abstract":"<div><p>The absorption, distribution, metabolism, excretion, and toxicity (ADMET) properties of drugs are critical to their efficacy and safety in clinical trials; however, traditional machine learning methods have limited generalization ability in ADMET prediction due to insufficient data. To address this issue, we developed DCPM-ADMET, an innovative pre-trained model with higher accuracy, whose architecture employs a two-component system, an XLNet-based module for capturing the deep semantics of molecular sequences, and a specialized RNN-based component (GRU) designed to encode global molecular property descriptors into high-dimensional latent representations for robust property extraction. By further incorporating ECFP fingerprints to capture local substructures, the model outperforms traditional methods and most pre-trained models in prediction accuracy on multiple benchmark datasets for molecular properties; additionally, we fine-tuned it on a self-constructed database containing 465,470 entries covering 97 ADMET properties, and by integrating these 97 prediction models and 36 computational properties, we further developed a free online ADMET prediction tool with 133 endpoints (available at http://admet.bioai-global.com/), which is designed to assist researchers in conducting comprehensive molecular ADMET predictions.</p><p><b>Scientific contribution</b></p><p>The development of DCPM-ADMET provides a robust and effective framework for molecular property prediction in computational pharmacology. Our architecture innovatively employs a dual-component system: an XLNet-based module for deep capture of molecular sequence semantics, a multi-task GRU module for joint SMILES translation and physicochemical property descriptor regression. Furthermore, we incorporate ECFP fingerprints to achieve exhaustive substructural feature encoding. Leveraging this multimodal fusion strategy, DCPM-ADMET exhibits superior predictive performance across diverse molecular property benchmark datasets, outperforming both traditional fingerprinting methods and state-of-the-art pre-trained models. Subsequently, we fine-tuned the model on a self-developed proprietary database—currently the largest of its category—comprising 465,470 entries that cover 97 ADMET endpoints (including 43 regression tasks, the highest number reported to date). By integrating the 97 resultant prediction models with 36 computed physicochemical properties, we have developed and made publicly available a free, high-throughput online ADMET prediction tool with 133 endpoints which is poised to serve as a novel and valuable alternative for guiding early-stage drug discovery and safety assessment.</p></div>","PeriodicalId":617,"journal":{"name":"Journal of Cheminformatics","volume":"18 1","pages":""},"PeriodicalIF":7.9,"publicationDate":"2026-06-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://link.springer.com/content/pdf/10.1186/s13321-026-01244-z.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148289674","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"化学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Learning a chemistry-aware latent space for molecular encoding and generation with a large-scale transformer variational autoencoder","authors":"Hugo Talibart,&nbsp;Dimitri Gilis","doi":"10.1186/s13321-026-01243-0","DOIUrl":"10.1186/s13321-026-01243-0","url":null,"abstract":"<div><p>Searching for molecules optimizing certain properties remains a key challenge due to the vastness of the chemical space, its discrete nature, and the limited availability of bioactivity data. One way to address these issues is to build a mapping from the chemical space to a continuous latent embedding space where efficient exploration and smooth interpolations become possible. Existing methods suffer from several limitations: failure to accurately reconstruct molecules, poorly structured latent space where whole regions cannot be decoded to valid molecular graphs or where distances fail to reflect chemical similarities, not to mention the unavailability of ready-to-use code or models trained on sufficiently large datasets, limiting their practical application. In this work, we provide a large-scale pre-trained Variational Autoencoder based on the Transformer architecture to convert small organic molecules to continuous fixed-size embeddings in a chemistry-aware structured latent space and back. With a <span>(97%)</span> reconstruction rate and a <span>(100%)</span> validity rate, the model provides a reliable mapping and is able to generate a high diversity of novel molecules. We introduce a training objective with a novel loss term explicitly enforcing embedding distances to reflect Tanimoto similarities between molecular fingerprints, leading to a smooth and well-structured latent space enabling seamless exploration and interpolation. We show that our model achieves solid performances on molecular optimization tasks compared to other generative models. We release an open-source, user-friendly implementation as a pip package and at github.com/3BioCompBio/chembed, along with pre-trained models, that can be readily used or fine-tuned for downstream drug design applications.</p></div>","PeriodicalId":617,"journal":{"name":"Journal of Cheminformatics","volume":"18 1","pages":""},"PeriodicalIF":7.9,"publicationDate":"2026-06-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://link.springer.com/content/pdf/10.1186/s13321-026-01243-0.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148262683","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"化学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Experimental dipole moment prediction with a progressive 2D–3D hybrid framework and twin-pair-based diagnostic evaluation","authors":"Sadettin Y. Ugurlu,&nbsp;Shan He","doi":"10.1186/s13321-026-01215-4","DOIUrl":"10.1186/s13321-026-01215-4","url":null,"abstract":"<p>Accurate prediction of molecular dipole moments (<span>(mu)</span>) is essential for modeling electrostatic interactions in solvation, molecular recognition, and materials. While modern 3D learning models can approach saturation on low-noise quantum-chemical benchmarks, performance often degrades on experimentally compiled datasets where measurement noise, stereochemical ambiguity, and conformer variability introduce a substantial gap between idealized theory and practical data. Here, we present a progressive 2D–3D hybrid framework for experimental dipole-moment prediction under heterogeneous labels. The framework is constructed in two stages. First, a strong 2D tabular CatBoost model based on SMILES-derived multi-view fingerprints is enriched with physicochemical, dipole-related, and conformer-dependent 3D shape descriptors, introducing a first level of 2D+3D integration and improving performance over the 2D representation alone. Second, this enriched tabular predictor is linearly fused with a geometry-aware 3D graph neural network operating on conformer-derived molecular graphs, yielding a further gain through complementary structural learning. On the experimental test set, the resulting framework achieves <span>(R^{2}=0.844)</span> with <span>(textrm{MAE}=0.399~textrm{D})</span>, outperforming a nonparametric fingerprint <i>k</i>-NN baseline (<span>(R^{2}=0.65)</span>) and remaining close to the similarity-conditioned diagnostic reference range estimated from twin-pair dispersion (<span>(R^{2}approx 0.86)</span>–0.91) under Tanimoto similarity thresholds of 0.95, 0.97, 0.98, and 0.99. </p><p>The main scientific contribution of this study is the development of an experimentally grounded progressive 2D–3D hybrid learning framework that integrates enriched tabular chemical representations, geometry-aware graph learning, and twin-pair-based diagnostic evaluation to predict dipole moments under heterogeneous experimental labels. To contextualize performance in this noise-affected experimental setting, we further use twin-pair analysis within a test-to-train similarity-stratified evaluation protocol, while baseline comparisons and component-wise interpretability analyses help clarify the roles of representation diversity and learner complementarity. Relative to three recently published dipole-prediction studies, the proposed framework remains competitively positioned, although these comparisons are intended only for cautious literature contextualization because the underlying datasets, label sources, and noise regimes are not directly matched. In contrast to prior studies focused mainly on lower-noise theoretical benchmarks or narrower molecular domains, the present work targets the comparatively underexplored Stenutz experimental compilation.</p>","PeriodicalId":617,"journal":{"name":"Journal of Cheminformatics","volume":"18 1","pages":""},"PeriodicalIF":7.9,"publicationDate":"2026-06-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://link.springer.com/content/pdf/10.1186/s13321-026-01215-4.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148248534","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"化学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"CRUSH—Cleavage Rules Using SMIRKS Heuristics: an enhanced molecular fragmentation algorithm","authors":"Edgar López-López,&nbsp;José L. Medina-Franco,&nbsp;Filip Miljković","doi":"10.1186/s13321-026-01239-w","DOIUrl":"10.1186/s13321-026-01239-w","url":null,"abstract":"<div><p>Molecular fragmentation methods are central to fragment-based discovery, virtual screening, and <i>de novo</i> molecular design, yet most existing approaches remain constrained by classical retrosynthetic rules that limit the scope and chemical space exploration beyond medicinal chemistry applications. Here we introduce CRUSH (Cleavage Rules Using SMIRKS Heuristics), an enhanced, chemistry-aware fragmentation algorithm that extends traditional retrosynthetic bond disconnection and building-block generation strategies to enable fine-grained and chemically meaningful molecular deconstruction. CRUSH adopts a flexible bond cleavage strategy in which a curated set of SMIRKS-based heuristics is applied exhaustively across all eligible bonds at each fragmentation step, enabling systematic exploration of multiple alternative disconnection pathways rather than committing to the first applicable retrosynthetic rule. Using a unified benchmarking framework across five chemically diverse datasets, including drug-like molecules, natural products, peptides, macrocycles, and alimentary compounds, we demonstrate that CRUSH consistently increases fragment yield and substantially expands chemical space coverage compared with the well-established methods. Notably, CRUSH-derived fragments access exclusive regions of chemical space that remain unexplored by existing approaches, a result that is consistent across both low- and high-resolution molecular fingerprint representations. Structurally, CRUSH generates smaller and less complex fragments, a profile particularly advantageous for fragment-based and diversity-driven workflows. Collectively, these results establish CRUSH as a robust and versatile molecular fragmentation framework that complements, and in exploratory contexts surpasses, current methods, providing a strong foundation for fragment-based chemoinformatics applications across multiple chemistry domains. As part of this work, implementation of the CRUSH fragmentation algorithm is made freely available at https://github.com/EdgL2/CRUSH.</p><p><b>Scientific contribution</b></p><p>This study introduces CRUSH, a chemistry-aware molecular fragmentation method that extends beyond traditional medicinal chemistry retrosynthetic rules to enable fine-grained and chemically meaningful bond disconnections. CRUSH consistently increases fragment yield and uniquely expands chemical space coverage across diverse datasets and molecular representations, generating smaller and less complex fragments well suited for fragment-based, exploratory, and <i>de novo</i> design workflows.</p></div>","PeriodicalId":617,"journal":{"name":"Journal of Cheminformatics","volume":"18 1","pages":""},"PeriodicalIF":7.9,"publicationDate":"2026-06-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://link.springer.com/content/pdf/10.1186/s13321-026-01239-w.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148248529","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"化学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Benchmarking molecular representations and machine learning algorithms for asymmetric catalysis: a palladium-catalysed decarboxylative asymmetric allylic alkylation case study","authors":"Eduardo Aguilar-Bejarano,&nbsp;Declan Galvin,&nbsp;David M. Rogers,&nbsp;Ender Özcan,&nbsp;Simon Woodward,&nbsp;Patrick J. Guiry,&nbsp;Grazziela Figueredo","doi":"10.1186/s13321-026-01236-z","DOIUrl":"10.1186/s13321-026-01236-z","url":null,"abstract":"&lt;div&gt;&lt;p&gt;Machine learning (ML) applied to metal-ligand asymmetric catalysis remains less explored compared to other applications, such as drug design or materials science. Current strategies frequently focus on augmenting pre-existing descriptors (e.g. those originally formulated for medicinal chemistry), the development of new bespoke steric and electronic descriptors, and the use of molecular fingerprints. Such method diversity, in the absence of user guidelines, makes selecting optimal ML tools to model new asymmetric catalysis problems challenging. This is exacerbated in early asymmetric catalysis metal-ligand development, where typically only limited data describing how ligands and substrates affect reaction stereoselectivity are available. Herein, we present a benchmarking pipeline for asymmetric catalysis ML studies and a comparative evaluation of reaction representations, including bespoke electronic and steric descriptors, fingerprints (CircuS and Morgan), physicochemical descriptors (from RDKit), machine-learnt representations (ChemBERTa, MolT5, and UniMol), and graph representations, paired with a range of ML algorithms (Support Vector Regressors, Random Forests, Extreme Gradient Boosting, Multilayer Perceptron, and GNNs). We use a database comprising only 103 early-development palladium-catalysed decarboxylative asymmetric allylic alkylation (DAAA) exemplars (three Trost-type ligands and 54 substrates) to predict reaction enantioselectivity, evaluating performance in low-data regimes and in extrapolative tasks under both random and substrate-scaffold splitting. An external validation dataset comprising 19 more recently developed reactions is used as a final evaluation. Across data regimes, Morgan fingerprints, ChemBERTa, and the bespoke V4 descriptors emerged as the most robust representations, with Random Forest the most consistent algorithm; on truly unseen substrates, an ensemble of GNNs gave the most accurate predictions. Importantly, scaffold-based splitting was found to estimate real-world extrapolation performance more reliably than random splitting. Post-hoc explainability analyses (SHAP and integrated gradients) revealed that the bespoke and fingerprint representations provided chemically meaningful insights into the substrate and catalyst features driving enantioselectivity, whereas embedding- and RDKit-based representations did not. Lastly, the methodology was validated on asymmetric hydrogenation and palladium-electrocatalysed C–H activation datasets, demonstrating its applicability to a wide range of asymmetric reactions.&lt;/p&gt;&lt;p&gt;&lt;b&gt;Scientific contribution&lt;/b&gt;&lt;/p&gt;&lt;p&gt;A curated dataset of 103 palladium-catalysed decarboxylative asymmetric allylic alkylation (DAAA) reactions, drawn consistently from Guiry group publications, is introduced together with an external validation set of 19 newer reactions. A bespoke, interpretable descriptor strategy that combines fragment-level steric (van der Waals volume) and electronic (Hammett-","PeriodicalId":617,"journal":{"name":"Journal of Cheminformatics","volume":"18 1","pages":""},"PeriodicalIF":7.9,"publicationDate":"2026-06-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://link.springer.com/content/pdf/10.1186/s13321-026-01236-z.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148238634","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"化学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"MutAIverse: an AI-powered, mechanism-backed platform for discovering novel DNA adducts and their precursor genotoxins","authors":"Shiva Satija,&nbsp;Sanjay Kumar Mohanty,&nbsp;Sachin B. Jorvekar,&nbsp;Anshul Verma,&nbsp;Saveena Solanki,&nbsp;Vishakha Gautam,&nbsp;Sakshi Arora,&nbsp;Aayushi Mittal,&nbsp;Arushi Sharma,&nbsp;Sonam Chauhan,&nbsp;Subhadeep Duari,&nbsp;Suvendu Kumar,&nbsp;Avdhesh Rai,&nbsp;Anupam Das,&nbsp;Kaberi Kakati,&nbsp;Kishore Das,&nbsp;Anupam Sarma,&nbsp;Juhi Tayal,&nbsp;Debarka Sengupta,&nbsp;Anurag Mehta,&nbsp;Roshan M. Borkar,&nbsp;Gaurav Ahuja","doi":"10.1186/s13321-026-01238-x","DOIUrl":"10.1186/s13321-026-01238-x","url":null,"abstract":"<div><p>Genotoxin exposure leads to DNA adduct formation, potentially causing mutations if unrepaired. Current DNA adductomics platforms or analytical workflows are limited by incomplete spectral libraries, reliance on experimentally validated adducts, limited cellular contexts, and inefficient computational methodologies. We introduce MutAIverse, an advanced AI-driven DNA adductomics analysis platform that overcomes these limitations by leveraging intracellular mechanistic modeling of genotoxin bioactivation to construct a comprehensive DNA adduct library and offers advanced spectral mapping. MutAIverse integrates experimentally validated and chemically valid putative DNA adducts, enabling interpretable retrograde tracking to parental genotoxins. Validation of MutAIverse against experimental MS/MS datasets demonstrated the detection of both known and novel adducts. Furthermore, application to in-house generated DNA adductomics data from tissue biopsies of smokeless tobacco-induced head and neck cancer patients revealed selective enrichment of both novel and known DNA adducts; a subset of them was validated using MS/MS analysis. Collectively, MutAIverse provides a robust, end-to-end, and interpretable platform for advanced DNA adductomics analysis.</p><p><b>Scientific Contribution</b></p><p>This study introduces MutAIverse, a generative AI‑powered platform that constructs a comprehensive, mechanism‑backed DNA adduct library from genotoxin metabolism simulations, enables interpretable retrograde tracing to parental genotoxins, and provides an end‑to‑end analytical workflow. By outperforming existing spectral libraries in coverage and experimental validation across multiple datasets, MutAIverse offers a robust tool for advancing DNA adductomics in toxicology and cancer research.</p></div>","PeriodicalId":617,"journal":{"name":"Journal of Cheminformatics","volume":"18 1","pages":""},"PeriodicalIF":7.9,"publicationDate":"2026-06-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://link.springer.com/content/pdf/10.1186/s13321-026-01238-x.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148222433","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"化学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Determination of bonding radii from small-molecule crystal structures","authors":"Eglė Šidlauskaitė,&nbsp;Andrius Merkys,&nbsp;Antanas Vaitkus,&nbsp;Algirdas Grybauskas,&nbsp;Saulius Gražulis","doi":"10.1186/s13321-026-01233-2","DOIUrl":"10.1186/s13321-026-01233-2","url":null,"abstract":"<div><p>X-ray crystallography rarely captures chemical bonding between atoms of a structure in question. Most of the time distance-based heuristics are applied to establish the pairs of bonded atoms. One class of such heuristics depends on a set of bonding radii that estimate the idealised size of each chemical element in a bond. This publication describes an unsupervised workflow for deriving a bonding radii set from crystal structure data in the Crystallography Open Database.  </p></div>","PeriodicalId":617,"journal":{"name":"Journal of Cheminformatics","volume":"18 1","pages":""},"PeriodicalIF":7.9,"publicationDate":"2026-06-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://link.springer.com/content/pdf/10.1186/s13321-026-01233-2.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148215769","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"化学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}