Establishment of an in silico-based framework for repurposing existing therapeutic agents to identify potential drugs targeting SARS-CoV

Authors

• Lin Ye
• Kel Nance
• Ste Kales
• Emi Lee
• Kha Shamim
• Val Labay
• Xin Hu

Abstract

SARS-CoV-2 is a positive single-strand RNA-based virus that has resulted in the loss of over 6.3 million lives since 2020. Developing drugs to combat this virus and its evolving variants cannot be overstated. In this paper, we developed an in-silico study-based framework focused on repurposing existing therapeutic agents to pinpoint potential drug candidates effective against COVID-19
In the initial step, we retrieved 256 drug-like bioactive molecules from the targeting the SARS coronavirus PLpro Protease. Utilizing the Max response and ccv2 parameters, we constructed models employing four distinct machine learning algorithms: Random Forest, Naïve Bayes, eXtreme Gradient Boosting, and Support Vector Machine. Model performance was assessed using the area under the receiver operating characteristic curve (AUC-ROC). Our comparative analysis revealed that the Random Forest (RF)-based QSAR model demonstrated superior predictive capability for the bioactivity of chemical compounds compared to XGBoost, SVM, and Naïve Bayes.
Additionally, employing the standard AC50 and efficacy we developed Random Forest-based regressor models. These models were utilized to screen the Genesis library for identifying suitable drug candidates against SARS-CoV-2.

Scientific Focus Area: Computational Biology

This page was last updated on Tuesday, August 6, 2024