Absorption, distribution, metabolism, excretion, and toxicity (ADMET) profiling is a major driver of drug success, controlling pharmacokinetic behavior, therapeutic effectiveness, and safety. Early identification of ADMET liabilities in pharmaceutical research and development (R&D) reduces attrition rates, optimizes candidate selection, and reduces the cost and time of clinical failures. In silico ADMET and toxicity prediction offers a cost-effective, ethically acceptable, and rapid alternative to traditional experimental methods, enabling data-driven decision-making in early stages of drug design. In this chapter, the ADMET paradigm and key pharmacokinetic and toxicological parameters used to assess drug-like properties are introduced. Computational methodologies are explored in detail, including rule-based approaches such as Lipinski’s Rule of Five and Veber’s rules, predictive modeling using quantitative structure–activity/property relationships (QSAR/QSPR), and sophisticated machine learning and deep learning algorithms for the capture of complex, nonlinear ADMET relationships. Public databases (e.g., ChEMBL, PubChem, Tox21) and commercial platforms (e.g., ADMET Predictor, SwissADME, pkCSM, Derek Nexus) are assessed for their utility in data curation and deployment of predictive models. Practical applications for absorption (e.g., permeability, P-gp efflux), distribution (volume of distribution, plasma protein binding), metabolism (CYP450-mediated biotransformation), and excretion pathways are reviewed. Toxicity modeling covers acute and chronic toxicity, organ-specific toxicity such as hepatotoxicity and cardiotoxicity, and long-term risks such as genotoxicity and carcinogenicity. A focused section provides an overview of ADME- and toxicity-specific databases and tools and their integration into early-stage drug development pipelines for rapid screening of high-risk candidates. Challenges such as heterogeneity of data quality, model interpretability, and translational gaps between computational predictions and in vivo outcomes are also discussed. Lastly, it discusses new trends such as AI-based multiparameter optimization, multi-omics data integration, and regulatory approval, putting in silico ADMET and toxicity prediction at the forefront of cutting-edge pharmaceutical innovation.

By Sohini Chakraborty, Hemanth Kumar Boyina, Raghavendra Mitta, Raghavendra Nayaka