QSAR Classification Models for Predicting 3CLPro-protease Inhibitor Activity

Abstract

The ongoing COVID-19 pandemic achieved a worldwide scale rapidly and has caused devastating casualties in terms of both human lives and in damage to the world economy. Several efforts for designing drugs and vaccines are underway across the globe. One of the potential early breakthroughs resulted due to the potential for repurposing existing drugs for COVID-19, specifically by drug modeling using computing power availability. Prediction of inhibition activity is a major step in such computation based drug discovery process. It is one of the virtual screening processes that throws light on particular molecules that may potential drug candidates. The subsequent stages in drug discovery are highly resource-intensive, during which a streamlined analysis of potential candidates can help in optimal design. Thus, the problem of predicting inhibition activity of compounds on proteins has attracted significant research interest. In this paper, an approach that employs quantitative structure-activity relationship (QSAR) modelling of SARS-CoV-3CLpro enzyme inhibitors for the development of activity classification model is proposed. The classification models predict SAR-CoV-3CLpro inhibitory activity for query compounds in the screening process using labels. Moreover, molecular docking analysis is performed using 3 FDA approved drugs that are being used as repurposed drugs for COVID-19 treatment. The best performing model with docking data (RMSD and Binding energy) of these 3 drugs were validated and the results obtained were promising.