Computational Toxicology最新文献

{"title":"Dual-site binding of BPA analogs to thyroid receptor alpha: A comprehensive computational study with relevance to autoimmunity and toxicological assessment","authors":"Rajesh Kumar Pathak,&nbsp;Alberto Fernández,&nbsp;Vikas Kumar","doi":"10.1016/j.comtox.2026.100447","DOIUrl":"10.1016/j.comtox.2026.100447","url":null,"abstract":"<div><div>Bisphenol A (BPA) is an endocrine disruptor that can bind to thyroid receptor alpha (TRα) and interfere with thyroid hormone signaling, a pathway implicated in thyroid autoimmunity in previous studies. As regulatory pressure drives the use of BPA analogs, concerns remain about whether these substitutes are safer than BPA. This study evaluated the interactions of BPA analogs with the ligand-binding domain (LBD) and allosteric site (AS) of TRα. Molecular docking was performed using AutoDock Vina and jdock, with triiodothyronine (T3) and BPA as reference ligands. Top-ranked complexes were further examined using 200 ns molecular dynamics simulations and MM-PBSA binding free-energy calculations. Density functional theory (DFT) was used to assess electronic properties, while absorption, distribution, metabolism, excretion, and toxicity (ADMET) profiling provided additional information on their predicted toxicological characteristics. Several analogs, particularly BisOPP-A and Bisphenol P, showed stronger predicted binding to TRα than BPA at both sites. BisOPP-A showed binding energies of −10.632 and − 10.89 kcal/mol at the LBD with AutoDock Vina and jdock, respectively, compared with −8.728 and − 8.85 kcal/mol for BPA. MM-PBSA further supported stronger BisOPP-A binding than BPA at the LBD (−41.96 vs −24.87 kcal/mol), together with measurable conformational changes in TRα. Overall, several analogs retained BPA-like characteristics while showing stronger predicted TRα interactions, indicating that structural substitution does not necessarily reduce thyroid receptor activity. These findings support further experimental validation of BPA analogs and integrated in silico approaches for early chemical risk assessment.</div></div>","PeriodicalId":37651,"journal":{"name":"Computational Toxicology","volume":"39 ","pages":"Article 100447"},"PeriodicalIF":3.1,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148951531","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Glyphosate as an acetylcholinesterase inhibitor: A multiscale computational study","authors":"Juan E. López-Cervantes,&nbsp;Carlos Z. Gómez-Castro,&nbsp;Luis A. Zárate-Hernández,&nbsp;Rosa L. Camacho-Mendoza,&nbsp;Simplicio Montiel-González,&nbsp;José M. Vásquez-Pérez,&nbsp;Marco A. Franco-Pérez,&nbsp;Julián Cruz-Borbolla","doi":"10.1016/j.comtox.2026.100446","DOIUrl":"10.1016/j.comtox.2026.100446","url":null,"abstract":"<div><div>This study evaluates glyphosate as a potential inhibitor of acetylcholinesterase (AChE), with classic organophosphates as a reference, using a multiscale computational strategy including molecular docking, conceptual density functional theory (CDFT) quantum theory of atoms in molecules (QTAIM) analysis, and quantum mechanics/molecular mechanics (QM/MM) simulations with umbrella sampling. Docking of glyphosate into the AChE catalytic active site, showed a suboptimal orientation of the phosphorus centre relative to the catalytic serine. A five-variable QSAR model built from CDFT-derived reactivity descriptors (ωB97X-D/6–311+G* level of theory) for 33 ChEMBL-derived organophosphates sharing an O=P–O fragment, gave a good internally validated fit (R² ≈ 0.92, Q²_LOO ≈ 0.89, Q²_LMO ≈ 0.88), with electrophilicity (<em>ω</em>) and the local electronic environment around phosphorus as the main contributors to activity. The model places glyphosate's predicted IC₅₀ in the millimolar range. The QM/MM umbrella sampling profile obtained with DFTB+/CHARMM36 for neutral non-zwitterionic glyphosate supports a two-step associative mechanism; direct comparison with the soman–AChE reference profile shows that glyphosate requires significantly higher activation energy for the phosphorylation step. The reaction for the neutral glyphosate modelled in this study is considered as a maximum-reactivity, lower-bound reference state; the additional thermodynamic cost of accessing this reactive protonated phosphonate state would further increase the effective barrier. Overall, docking, QSAR, and QM/MM independently converge on the conclusion that, although covalent phosphorylation of AChE by glyphosate is mechanistically conceivable, it is inefficient under physiological conditions. These results provide a computational rationale for the poor inhibitory activity already reported experimentally.</div></div>","PeriodicalId":37651,"journal":{"name":"Computational Toxicology","volume":"39 ","pages":"Article 100446"},"PeriodicalIF":3.1,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148951532","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Machine-readable immune ontogeny maps for AOP-anchoring: A critical review","authors":"Rajesh Pamanji","doi":"10.1016/j.comtox.2026.100449","DOIUrl":"10.1016/j.comtox.2026.100449","url":null,"abstract":"<div><div>The adverse outcome pathway framework has become a central organizing structure in regulatory and predictive toxicology, and a well-resourced international effort is now underway to build a machine-readable map of human immune system development for exactly this purpose. This review critically examines that emerging effort and the surrounding technical landscape, asking what a machine-readable immune ontogeny map needs to contain, how it should be represented, and how far current and proposed resources actually get toward supporting adverse outcome pathway key events as computable, age-indexed biological states. We describe the Human Immune System Development Map, a systems-biology network representation built on disease-map conventions, situate it within the wider Partnership for the Assessment of Risks from Chemicals developmental immunotoxicity initiative, and compare its representational choices against an alternative built on standardized biomedical ontologies for cell type, anatomy, and developmental stage. We survey human and zebrafish immune ontogeny atlases, the ontology infrastructure available for cross-species annotation, and emerging large language model-based tools for adverse outcome pathway curation, and we identify specific respects in which the existing map is human-only, based on a graphical systems-biology representation rather than formally ontology-tagged, and not yet interoperable with the semantic web representation of the adverse outcome pathway wiki. We propose a complementary, ontology-tagged and cross-species extension architecture designed to interoperate with, rather than duplicate, the existing systems-biology map, and we identify the curation, versioning, and validation obstacles that stand between current resources and routine regulatory use. The review concludes that the foundational infrastructure for this integration is being actively built, but that closing the gap between systems-biology network maps and formally ontology-anchored, cross-species adverse outcome pathway key events remains substantially unfinished work.</div></div>","PeriodicalId":37651,"journal":{"name":"Computational Toxicology","volume":"39 ","pages":"Article 100449"},"PeriodicalIF":3.1,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148951533","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A novel reuse of data from lab mouse models to inform chemical impacts on the genetics of wild mouse populations","authors":"Daniel Guignard,&nbsp;Maria Büttner,&nbsp;Tiffany Scholier,&nbsp;Marissa B. Kosnik","doi":"10.1016/j.comtox.2026.100419","DOIUrl":"10.1016/j.comtox.2026.100419","url":null,"abstract":"<div><div>Pollution is a major but under-evaluated driver of biodiversity loss. While methods are being developed to predict genotype-environment (GxE) interactions for humans, such approaches are under-developed for non-human populations. We built a computational approach to predict GxE interactions for <em>Mus musculus</em>, an ecologically relevant species. We integrate disconnected datasets to form Chemical – Pathway – Gene – Variant – Phenotype linkages for three chemical classes (pharmaceuticals, pesticides, and cosmetics) with two sets of genetic variants: from primarily lab mice in Ensembl (&gt;15,000,000 variants) and from wild mice (&gt;50,000 variants) to identify genes and pathways that may underlie genetic susceptibility or adaptation to chemicals in wild mouse populations. By comparing the Ensembl and wild mouse linkages and the three chemical classes, we 1) characterize pathways implicated in GxE interactions (e.g., PI3K-Akt signaling pathway), 2) identify genes with high variability in the wild mouse linkages that may be important in GxE interactions (e.g., IL6, NFKB1), and 3) highlight genes from Ensembl linkages in the same pathways as the wild linkages, but absent from the wild mouse data (e.g., TLR4, TNF). While some variants implicated in Ensembl may differ from what would be in wild populations, we propose that the genes may still be important given that they are in the same pathways implicated in the wild linkages. Future studies can prioritize the genes and pathways we highlight to characterize genetic susceptibility or adaptation to chemicals for wild mouse populations and expand our approach of repurposing lab data to inform GxE interactions in other ecologically relevant species.</div></div>","PeriodicalId":37651,"journal":{"name":"Computational Toxicology","volume":"38 ","pages":"Article 100419"},"PeriodicalIF":2.9,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148183690","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"ProtoAquaTox: versatile aquatic toxicity prediction software across multiple aquatic organisms and experimental conditions","authors":"Pravin Ambure ,&nbsp;Eva Serrano-Candelas ,&nbsp;Oscar Andreu-Sánchez ,&nbsp;María Luisa Fernández-Cruz ,&nbsp;Jyotsna Bhat-Ambure ,&nbsp;Rafael Gozalbes","doi":"10.1016/j.comtox.2026.100420","DOIUrl":"10.1016/j.comtox.2026.100420","url":null,"abstract":"<div><div>Aquatic toxicity testing, which traditionally involves in-vivo experiments on organisms across three trophic levels (algae, invertebrates, and fish), is both resource-intensive and ethically challenging. As an alternative, <em>in-silico</em> methods such as Quantitative Structure-Toxicity Relationship (QSTR) modeling have emerged as promising tools for predicting chemical toxicity.</div><div>In this research, we present ProtoAquaTox®, a software platform developed to predict acute and chronic aquatic toxicity using three multi-tasking QSTR models. The platform employs multi-tasking QSTR models built on a dataset of 2119 unique chemicals and 11,860 toxicity values, representing acute and chronic toxicity data for different trophic levels, namely, fish, crustaceans, and algae. These models encompass multiple test organisms (total 87 species) from all three trophic levels and are designed to account for variations in experimental parameters such as test organisms, media type, exposure type, duration, etc. Employing the random forest machine learning technique, we developed three binary classification models (Models1–3) that categorize chemicals into distinct toxicity classes, achieving optimal performance based on validation metrics. Our multi-tasking models demonstrate superior predictive performance, achieving accuracies between 0.85 and 0.90, precision values ranging from 0.85 to 0.91, MCC scores between 0.68 and 0.79, and AUC values in the range of 0.92 to 0.96 when evaluated on an external test set. Experimental validation against test organisms has demonstrated the models' efficacy, with areas for further refinement identified. ProtoAquaTox® provides an intuitive interface, enabling users to predict the environmental risks posed by diverse chemicals, thereby aiding hazard mitigation and contributing to the protection of aquatic ecosystems. The platform is freely available through the Chemopredictionsuite web platform, making it a valuable tool for environmental risk assessment.</div></div>","PeriodicalId":37651,"journal":{"name":"Computational Toxicology","volume":"38 ","pages":"Article 100420"},"PeriodicalIF":2.9,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148183691","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Physiologically based kinetics and toxicodynamic (PBK-TD) modelling in zebrafish embryos: A novel approach to assess endocrine disrupting chemicals acting on brain aromatase","authors":"Rémy Beaudouin ,&nbsp;Florian Geffroy ,&nbsp;Karyn Le Menach ,&nbsp;Hélène Budzinski ,&nbsp;Cleo Bodin ,&nbsp;Romain Dussauçoy ,&nbsp;Benjamin Piccini ,&nbsp;Edith Chadili ,&nbsp;Nathalie Hinfray ,&nbsp;François Brion ,&nbsp;Pierre-André Billat","doi":"10.1016/j.comtox.2026.100423","DOIUrl":"10.1016/j.comtox.2026.100423","url":null,"abstract":"<div><div>Endocrine-disrupting chemicals (EDCs) can interfere with neurodevelopment through modulation of estrogen signaling pathways, including brain aromatase (<em>cyp19a1b</em>) expression. However, linking external exposure to internal dose and biological effects remains a key challenge for risk assessment. In this study, we developed a physiologically based kinetic–toxicodynamic (PBK-TD) model to describe the time- and concentration-dependent induction of brain aromatase in zebrafish embryos. Using EASZY assay protocol, transgenic tg(<em>cyp19a1b</em>:GFP) embryos were exposed to three model estrogenic compounds, 17α-ethinylestradiol (EE2), bisphenol A (BPA), and bisphenol F (BPF), from 0 to 96 h post-fertilization. GFP fluorescence and RT-qPCR measurements were used to characterize temporal dynamics of brain aromatase expression. A previously established PBK model was adapted to predict internal concentrations, particularly in the brain, and coupled to an indirect response TD model describing aromatase induction. The PBK model adequately predicted internal concentrations for all compounds (majority within 3-fold error). The PBK-TD model successfully captured the onset of inducibility (∼44 hpf) and reproduced the full time-course of GFP responses. Internal dose–response analysis revealed substantial differences in potency between compounds, with EE2 showing markedly higher potency than bisphenols. Importantly, relative potency rankings differed when based on internal versus external concentrations, including an inversion between BPA and BPF. Overall, this work demonstrates that combining EASZY assay and PBK-TD modelling provides a more mechanistically relevant assessment of endocrine activity, improved interpretation of zebrafish embryo assays and is first step toward for quantitative in vitro–in vivo extrapolation and qAOP development.</div></div>","PeriodicalId":37651,"journal":{"name":"Computational Toxicology","volume":"38 ","pages":"Article 100423"},"PeriodicalIF":2.9,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148183833","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Corrigendum to “Variation of the amino acid at position 107 is a major contributor to species-specific sensitivity to PPO inhibitors” [Comput. Toxicol. 38 (2026) 100415]","authors":"Michael Betz,&nbsp;Anna Berteotti,&nbsp;Aimone Porri,&nbsp;Matthias Witschel,&nbsp;Tobias Seiser,&nbsp;Thomas Seitz,&nbsp;Markus Frericks,&nbsp;Brandy Riffle","doi":"10.1016/j.comtox.2026.100416","DOIUrl":"10.1016/j.comtox.2026.100416","url":null,"abstract":"","PeriodicalId":37651,"journal":{"name":"Computational Toxicology","volume":"38 ","pages":"Article 100416"},"PeriodicalIF":3.1,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148523553","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Integrating machine learning with aggregate risk models to assess acute myeloid leukemia from chemical mixture exposure","authors":"Seshill Real Armas ,&nbsp;Mesha Ajaz ,&nbsp;Daniela Galatro ,&nbsp;Alessia Di Nardo ,&nbsp;David A. Romero ,&nbsp;Rosario Trigo-Ferre ,&nbsp;Melanie Jeffrey ,&nbsp;Maria Jacome ,&nbsp;Vincenzo Costanzo-Alvarez ,&nbsp;Jason Bazylak ,&nbsp;Cristina H. Amon","doi":"10.1016/j.comtox.2026.100425","DOIUrl":"10.1016/j.comtox.2026.100425","url":null,"abstract":"<div><div>Traditional aggregate risk models rely on the availability of toxicological parameters for each substance in the mixture and assume a linear relationship between the substances. However, these models are limited by incomplete toxicological data and the potential for non-linear (antagonistic or synergistic) interactions among substances. To address these limitations, this work presents a framework for characterizing and interpreting toxicological data to support health risk prediction. Toxicological parameters were obtained from the United States Environmental Protection Agency's Integrated Risk Information System database, and exposure data from an artificially generated dataset based on the properties of exposure data published by Heck et al. The latter dataset is highly imbalanced, with a ratio of negative to positive individuals of 0.0024 of total samples, thus posing modelling challenges. Specifically, we apply Principal Component Analysis (PCA) to infer relationships among exposure substances, identify clusters in the principal component space to estimate missing toxicological parameters, and use the resulting PCA loadings to cross-check linear aggregation weights learned via gradient descent. We then extend the linear risk aggregation model to include pairwise interaction terms to capture non-linear dependencies and compare its interpretability and predictive performance against three tree-based models: Random Forest (RF), Balanced Random Forest (BRF), and eXtreme Gradient Boosting (XGBoost). Our results show that PCA reveals groupings of high-risk pollutants that align with both the established literature and the gradient-learned linear aggregation weights. Interaction-based extensions to the linear model produced unstable and uninterpretable results, likely due to extreme data imbalance and the limitations of using pairwise exposure products to represent interactions. Among tree-based methods, BRF achieved the best performance on imbalanced data, and using a minority oversampling technique to balance the data was insufficient to achieve strong predictive performance on the extremely imbalanced datasets typically encountered in toxicological risk prediction applications.</div></div>","PeriodicalId":37651,"journal":{"name":"Computational Toxicology","volume":"38 ","pages":"Article 100425"},"PeriodicalIF":2.9,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148241618","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Computational approaches to adverse outcome pathway networks: advances in construction, quantification, and regulatory decision support","authors":"Abdallah Abou Hajal,&nbsp;Ahmad Z.Al Meslamani","doi":"10.1016/j.comtox.2026.100429","DOIUrl":"10.1016/j.comtox.2026.100429","url":null,"abstract":"<div><div>Adverse outcome pathways (AOPs) provide a structured causal framework from molecular initiating events to adverse outcomes, yet most AOP applications remain qualitative and lack the quantitative, uncertainty-aware predictions needed for dose–response extrapolation—especially when biology operates as interacting AOP networks. This review outlines how computational techniques, including machine learning (ML), deep learning (DL), natural language processing (NLP), and knowledge-graph workflows, construct, enrich, quantify, and validate AOP networks. It further explains how these networks connect to omics and new approach methodologies (NAMs) for human-predictive toxicology. We searched PubMed, Embase, Scopus, and Web of Science (January 2010–December 2025), supplemented by backward/forward citation chasing and targeted retrieval from the OECD AOP Knowledge Base and AOP-Wiki. Computational approaches now allow scalable AOP network building from heterogeneous data, automate evidence extraction from the literature, and support quantitative modeling of key event relationships. When anchored to NAM data, these computationally augmented networks can mechanistically interpret in vitro and omics findings. For routine regulatory use, Computational AOP networks are best deployed as auditable decision-support tools for screening and prioritization. Their reliable application will require robust quantitative key-event correlations, well-defined applicability domains, explicit uncertainty quantification, and systematic toxicokinetic integration to translate in vitro signals to human exposure contexts.</div></div>","PeriodicalId":37651,"journal":{"name":"Computational Toxicology","volume":"38 ","pages":"Article 100429"},"PeriodicalIF":2.9,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148241616","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Activity cliff-guided feature selection for improved toxicity prediction using machine learning","authors":"Chenjia Li ,&nbsp;Jiangxue Qian ,&nbsp;Yumian Zhou,&nbsp;Shou-Lin Wang,&nbsp;Chao Chen","doi":"10.1016/j.comtox.2026.100428","DOIUrl":"10.1016/j.comtox.2026.100428","url":null,"abstract":"<div><div>Accurate in silico toxicity prediction is essential for chemical safety assessment and computational toxicology. This study presents a novel Cliff method that leverages structure-activity relationship (SAR) information from activity cliff pairs to enhance machine learning models for toxicity prediction. We analyzed 249 bioassays from the Tox21 dataset and identified activity cliff pairs based on Tanimoto similarity and activity difference analysis. A total of 1613 molecular descriptors were calculated using Python Mordred, and single-feature decision tree models evaluated the predictive power of each descriptor across the cliff pair subsets of each bioassay dataset. Through construction and evaluation of multiple machine learning models across all datasets, we demonstrated that Cliff outperformed conventional feature engineering approaches, including variance thresholding, correlation analysis, RFE, and LASSO regression. Cliff achieved a superior median AUC of 0.736, with particularly notable improvements observed in non-tree-based machine learning models such as logistic regression and SVM (average AUC gains up to 0.071). Further analysis revealed that cliff pairs showed significantly higher descriptor value differences than non-cliff pairs for seven key descriptors, including d_ATS8Z, d_AATSC1d, d_AATSC5c, d_SLogP, d_ATSC2d, d_WPath, and d_ATS8m. These findings highlight the potential of activity cliff-guided feature selection as a valuable computational tool for improving the predictive accuracy and interpretability of in silico toxicity models.</div></div>","PeriodicalId":37651,"journal":{"name":"Computational Toxicology","volume":"38 ","pages":"Article 100428"},"PeriodicalIF":2.9,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148241619","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}