Predictive QSAR modeling: Methods and applications in drug discovery and chemical risk assessment

Abstract

Quantitative structure-activity relationship (QSAR) modeling is the major chemin- formatics approach to exploring and exploiting the dependency of chemical, biological, toxicological, or other types of activities or properties on their molecular features. QSAR modeling has been traditionally used as a lead optimization approach in drug discovery research. However, in recent years QSAR modeling found broader applications in hit and lead discovery by the means of virtual screening as well as in the area of drug-like property prediction, and chemical risk assessment. These developments have been enabled by the improved protocols for model development and most importantly, model validation that focus on developing models with independently validated external prediction power. This chapter reviews the predictive QSAR modeling workflow developed in this laboratory that incorporates rigorous procedures for QSAR model development, validation, and application to virtual screening. It also provides several examples of the workflow application to the identification of experimentally confirmed hit compounds as well as to chemical toxicity modeling. We believe that methods and applications considered in this chapter will be of interest and value to researchers working in the field of computational drug discovery and environmental chemical risk assessment.