Quantitative Toxicophore Discovery through Interpretable Deep Learning: A Framework for Mechanism-Guided Toxicity Mitigation in Drug Safety Profiling
ACS Omega · 06.11.2026
DOI: 10.1021/acsomega.6c08546 ↗Kurz gefasst
Fachartikel in der Zeitschrift ACS Omega, online erschienen am 06.11.2026. Verfasst von Jing Gao und 2 weiteren Autor:innen. Die Arbeit nennt 31 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-NC-ND 4.0).
Abstract (Originalsprache)
Unanticipated toxicity accounts for ∼30% of clinical-phase drug failures, yet despite advances in computational toxicity prediction, predictive models typically offer scores without identifying the structural culprit. Here, we introduce an interpretable deep-learning framework that couples fingerprint-descriptor fusion with task-adaptive attention and a SHAP-guided bioisosteric ablation pipeline. Trained on 9,315 compounds spanning 14 end points (Tox21 + ClinTox) under scaffold-stratified split, the attention-fusion model achieves mean test ROC-AUC of 0.739 (Tox21) and 0.953 (ClinTox). The critical advance is a structure-guided validation protocol: predictive intervention ablation─computationally replacing toxicophores with bioisosteric equivalents─demonstrates that the model’s predictions are predictively dependent on specific substructures. Halogen → methyl substitution drops predicted toxicity from 0.659 to 0.101 (paired t = 22.66, p = 6.32 × 10–10, n = 11; Cohen’s d = 6.82), providing the primary structure-guided evidence. Cross-end point SHAP correlation reaches 0.333 (MTL) and 0.334 (STL; Welch’s p = 0.82), indicating that both paradigms extract coherent mechanistic structure above the label baseline (0.172). SMARTS-based toxicophore alignment identifies a strong positive logit differential for aniline on the clinically relevant CT_TOX end point (+0.76), and predictive-intervention ablation confirms model dependence on the identified substructures (halogen → methyl: n = 11, p < 10–9; aniline → methoxy: n = 2, directional check only). External validation on 1,427 SIDER adverse-drug-reaction compounds reveals significant toxicophore enrichment (aniline: 12.6% vs 0%, Fisher exact p = 4.5 × 10–12; halogenated aromatic: 29.8% vs 0%, p = 1.6 × 10–29), remaining significant after controlling for molecular weight and Log P. Three case studies illustrate how SHAP-guided attribution translates into actionable medicinal-chemistry mitigation strategies.
Zitieren
APA
Gao, J., Lin, D., & Hu, J. (2026). Quantitative Toxicophore Discovery through Interpretable Deep Learning: A Framework for Mechanism-Guided Toxicity Mitigation in Drug Safety Profiling. ACS Omega. https://doi.org/10.1021/acsomega.6c08546
Vancouver
Gao J, Lin D, Hu J. Quantitative Toxicophore Discovery through Interpretable Deep Learning: A Framework for Mechanism-Guided Toxicity Mitigation in Drug Safety Profiling. ACS Omega. 2026. doi:10.1021/acsomega.6c08546
BibTeX
@article{Gao2026,
author = {Gao, Jing and Lin, Dong and Hu, Jianfeng},
title = {Quantitative Toxicophore Discovery through Interpretable Deep Learning: A Framework for Mechanism-Guided Toxicity Mitigation in Drug Safety Profiling},
journal = {ACS Omega},
year = {2026},
doi = {10.1021/acsomega.6c08546}
}
EUVIMED