Advances in the Application of In Silico ADMET Models – An Industry Perspective
Wenyi Wang, Fjodor Melnikov, Joe Napoli, Prashant V. Desai · 2024
Small molecule drug discovery involves extensive evaluation of potencies, molecular physicochemical properties, and pharmacokinetic (PK) properties related to absorption, distribution, metabolism, excretion, and toxicities (ADMET). In silico ADMET tools, utilizing diverse computational techniques, have been demonstrated to significantly improve the efficiency of drug discovery, reduce costs, and provide safer and more efficacious drugs to patients faster. In this chapter, we provide an overview of various in silico models routinely used during the drug discovery process to design and prioritize compounds with optimal ADMET and PK properties. Given that quantitative structure–activity relationship (QSAR) models are the most commonly used in silico models for predicting ADMET properties, this chapter provides an in-depth description of how these models are built, evaluated, and applied. The majority of the examples are taken from industrial settings describing some of the best practices with an emphasis on prospective application. Other varieties of in silico ADMET models, such as the mechanistic models and predictive models for site of metabolism, are also briefly discussed.