Towards the discovery of molecules with anti-COVID-19 activity: Relationships between screening and docking results

Abstract

The study presents the results of a combined approach to the theoretical description of potential antiviral activity against COVID-19. We found that pharmacophore screening based on limited experimental data on "protein-ligand" binding complexes might have low predictive ability. Therefore, in this study, we build a model based on the statistical description of QSAR for data obtained from docking which serves as a basis for adequate prediction of ligand activity. We use the logistic regression to construct the predictive model for the main protease Mpro inhibitors.

Keywords

• qsar
• docking
• pharmacophore
• logistic regression