QSAR in Toxicology: Methodological Foundations and Computational Workflows

This methodological guide provides a comprehensive overview of data-driven approaches to Quantitative Structure–Activity Relationship (QSAR) modeling in toxicology. It introduces fundamental and advanced methods for data curation, descriptor generation, unsupervised and supervised learning, validation, applicability domain (AD) analysis, and model interpretation. Emphasis is placed on reproducible research, mechanistic interpretation, and regulatory alignment. The material integrates theoretical background with practical exercises and reproducible workflows implemented in modern computational environments.