Drug Repurposing Targeting COVID-19 3CL Protease using Molecular Docking and Machine Learning Regression Approach

Aug 5, 2026

Upload by Logic

The COVID-19 pandemic created an urgent need for rapid therapeutic discovery. This study screens 5903 FDA-approved drugs from the Zinc database for repurposing against the SARS-CoV-2 main protease 3CL. Molecular docking is performed with AutoDock-Vina to assess binding efficacy, while binding affinities are modeled using QSAR-focused machine learning regression methods including decision tree, extra trees, MLP, KNN, XGBoost, and gradient boosting. Decision Tree Regression improves R2 and RMSE and supports shortlisting 13 candidates with binding energies from -15.1 to -12.7 kcal/mol, followed by physiochemical property analysis.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages30

Reading Time1 h 16 min

AccessFree to download

Abstract

1 Introduction

COVID-19 and lack of specific therapeutics

Role of SARS-CoV-2 3CLpro as a target

Computational drug discovery: docking and machine learning

Drug repurposing strategy for COVID-19

Drug Repurposing Targeting COVID-19 3CL Protease using Molecular Docking and Machine Learning Regression Approach Imra Aqeel 1, Abdul Majid 1 1 Biomedical Informatics Research Lab, Department of Computer & Information Sciences, Pakistan Institute of Engineering & Applied Sciences, Nilore, Islamabad 45650, Pakistan; [imraaqeel@pieas.edu.pk](imraaqeel@pieas.edu.pk) Abstract: The COVID-19 pandemic has created a global health crisis, driving the need for the rapid identification of potential therapeutics. In this study, we used the Zinc database to screen the worldapproved including FDA-approved 5903 drugs for repurposing as potential COVID-19 treatments targeting the main protease 3CL of SARS-CoV-2. We performed molecular docking using Autodock-Vina to check the efficacy of drug molecules. To enhance the efficiency of drug repurposing approach, we modeled the binding affinities using several machine learning regression approaches for QSAR modeling such as decision tree, extra trees, MLP, KNN, XGBoost, and gradient boosting. The computational results demonstrated that Decision Tree Regression (DTR) model has improved statistical measures of R2 and RMSE. These simulated results helped to identify drugs with high binding affinity and favorable binding energies. From the statistical analysis, we shortlisted 13 promising drugs with their respective Zinc IDs (ZINC000003873365, ZINC000085432544, ZINC000203757351, ZINC000085536956, ZINC000085536990, ZINC000008214470, ZINC000261494640, ZINC000169344691, ZINC000094303244, ZINC000095618608, ZINC000095618689, ZINC000095618743, and ZINC000253684767) within the range of-15.1 kcal/mol to -12.7 kcal/mol. Further, we analyzed the physiochemical properties of these selected drugs with respect to their best binding interaction to specific target protease. Our study has provided an efficient framework for drug repurposing against COVID-19. This highlights the potential of combining molecular docking with machine learning regression approaches to accelerate the identification of potential therapeutic candidates. Keywords: CoVID-19; main protease 3CL; drug repurposing; QSAR model; binding affinity; molecular docking 1 Introduction The COVID-19 pandemic has presented an unprecedented global health crisis, with over 687 million confirmed cases and over 6.8 million deaths worldwide as of May 2023 according to [https://www.worldometers.info/coronavirus/.](https://www.worldometers.info/coronavirus/. Currently)[ Currently](https://www.worldometers.info/coronavirus/. Currently), there is no specific drug available to treat COVID-19, and the development of effective therapies has become a priority for researchers globally [1]. COVID-19 is caused by the severe acute respiratory syndrome coronavirus 2 (SARSCoV-2), a positive-sense single-stranded RNA virus that primarily infects the respiratory tract of humans [2] . The entry of the virus into host cells occurs when the spike protein binds to the ACE2 receptor on the surface of human cells, and then it utilizes the host's cellular machinery to replicate and spread throughout the body. Figure 1 depicts the biological structure of coronavirus with its structural proteins along with the crystal structure of viral 3CL protease. Figure 1: Structure of coronavirus along with crystal structure of SARS 3CL protease To support viral replication, SARS-CoV-2 uses various viral proteins, among which the main protease 3CLpro (also called Main protease Mpro) plays a crucial role in cleaving the viral polyproteins into functional non-structural proteins necessary for viral replication. As a result of its significance in the viral life cycle, 3CLpro has become a potential target for the development of antiviral therapies for COVID-19. The catalytic dyad of His41 and Cys145 in the homodimeric cysteine protease 3CL protease makes it an attractive candidate for the development of protease inhibitors [3] . Several studies have reported the successful identifica

FAQs

What is the main target for drug repurposing in this study?

The study targets the SARS-CoV-2 main protease 3CL (3CLpro/Mpro), which is essential for processing viral polyproteins during replication.

How are drug molecules evaluated in the workflow?

Drugs are first screened from the Zinc database and then assessed using molecular docking with AutoDock-Vina to estimate binding efficacy.

Which machine learning approach performed best for binding affinity prediction?

Decision Tree Regression (DTR) showed improved statistical measures, including higher R2 and lower RMSE, supporting better selection of promising candidates.

Aug 5, 2026

Upload by Logic