Statistical learning approaches for predicting pharmacological properties of pharmaceutical agents

Abstract

Drug development is aimed at the finding of therapeutic agents that possess desirable pharmacological properties. Historically, inappropriate pharmacological properties have been the primary reasons for the failure of drug candidates in the later stages of drug development. Thus, tools for predicting pharmacological properties in earlier drug design stages are needed. As part of the efforts, computational approaches have been explored for predicting various pharmacological properties of pharmaceutical agents. This work aims to study the applicability of statistical learning methods (SLMs) to classify compounds from diverse structures into different pharmacological properties. Our results show that SLMs can improve the quality of the pharmacological properties prediction models by using enlarged and more diverse groups of compounds. Recursive feature elimination is able to identify a group of relevant molecular descriptors that reflect the pharmacological properties. Moreover, SLMs are found to be useful for developing clinically important but insufficiently explored pharmacological properties prediction systems.