Unveiling the Inhibitory Potential of Gingerol Substructures against SARS- CoV-2 RBD: An Integrated Ensemble Learning and In Silico Screening Approach

Abstract

Abstract In the relentless quest for effective treatments against SARS-CoV-2, extensive exploration of potential inhibitors has been underway. In this study, we present an integrated approach combining machine learning and in silico screening to identify promising inhibitors for the SARS-CoV-2 receptor-binding domain (RBD). We harnessed a dataset of Vina scores for 988 gingerol substructures, employing Random Forest (RF) regression as the optimal model to predict Vina scores accurately (R² = 0.77). Virtual screening, both through RF predictions and PyRx, consistently highlighted 14 molecules with inhibitory potential. Pharmacokinetic evaluation, aided by the Bioavailability Radar and a BOILED-Egg simulation, further refined the selection of four leads-G4, G5, G11 and G13 with human intestinal absorption, out of which the P-gp non substrate G13 (PubChem CID: 135196841) can be act as a promising candidate. Molecular docking, molecular dynamics simulations, and Density Functional Theory (DFT) calculations validated the stability and interactions of this compound with the SARS-CoV-2 RBD. Our study offers a streamlined methodology for identifying potential inhibitor, paving the way for further experimental validation.