Scholarlink

<p>The increasing demand for accurate toxicity assessment and to minimise or eliminate the use of animal testing has accelerated the development of numerous computational models, AI/ML models, and online resources that support research in computational toxicology. This review addresses toxicity prediction and chemical safety evaluation, focusing on computational models and data coverage, Molecular Descriptors, QSAR models, AI/ML-based approaches, Explainable AI, predictive methodologies, regulatory relevance, and accessibility. This collectively enables the identification, prediction, and analysis of chemical toxicity across various biological endpoints. In addition, the review highlights AI/ML tools for predicting toxicity endpoints, such as neurotoxicity, hepatotoxicity, cardiotoxicity, genotoxicity, and environmental toxicity. Regulatory limitations vary significantly among countries and jurisdictions, exhibiting a marked absence of convergence. Current debates regarding regulatory norms focus on achieving global conformity. Regulatory adaptability is the key as AI evolves rapidly. The promotion of AI/ML tool integration and interoperable frameworks could substantially enhance the future of predictive toxicology.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

AI/ML-based computational models for toxicity prediction

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年1月12日

查看全文
  • Sushmita Barua,
  • Badhrinarayanan Balaji,
  • Seetharaman Balaji

摘要

The increasing demand for accurate toxicity assessment and to minimise or eliminate the use of animal testing has accelerated the development of numerous computational models, AI/ML models, and online resources that support research in computational toxicology. This review addresses toxicity prediction and chemical safety evaluation, focusing on computational models and data coverage, Molecular Descriptors, QSAR models, AI/ML-based approaches, Explainable AI, predictive methodologies, regulatory relevance, and accessibility. This collectively enables the identification, prediction, and analysis of chemical toxicity across various biological endpoints. In addition, the review highlights AI/ML tools for predicting toxicity endpoints, such as neurotoxicity, hepatotoxicity, cardiotoxicity, genotoxicity, and environmental toxicity. Regulatory limitations vary significantly among countries and jurisdictions, exhibiting a marked absence of convergence. Current debates regarding regulatory norms focus on achieving global conformity. Regulatory adaptability is the key as AI evolves rapidly. The promotion of AI/ML tool integration and interoperable frameworks could substantially enhance the future of predictive toxicology.