Integrating Molecular Dynamics, Molecular Docking, and Machine Learning for Predicting SARS-CoV-2 Papain-like...

Area of Science:

• Computational biology
• Drug discovery
• Virology

Background:

• Coronavirus disease 2019 (COVID-19) has caused significant global health and economic disruption.
• Effective antiviral therapies for COVID-19 are limited, especially those targeting the SARS-CoV-2 papain-like protease (PLpro), crucial for viral replication and immune evasion.

Purpose of the Study:

• To identify potential SARS-CoV-2 PLpro inhibitors through drug repurposing using computational methods.
• To accelerate the discovery of new antiviral treatments for COVID-19.

Main Methods:

• Combined machine learning, molecular dynamics simulations, and molecular docking to screen FDA-approved drugs.
• Performed long-timescale molecular dynamics simulations on PLpro-ligand complexes.
• Utilized a random forest model trained on docking scores for prediction, achieving 76.4% accuracy.

Main Results:

• Identified five FDA-approved drugs as promising candidates for COVID-19 treatment by repurposing.
• The machine learning model demonstrated high accuracy in predicting potential PLpro inhibitors.
• Successfully filtered drug candidates based on prediction confidence and applicability domain.

Conclusions:

• Integrating computational modeling and machine learning is effective for accelerating drug repurposing against emerging viral targets like SARS-CoV-2 PLpro.
• The identified drugs represent potential new therapeutic options for COVID-19 treatment.
• This approach can expedite the development of antiviral therapies for future pandemics.