Shape-based Machine Learning Models for the Potential Novel COVID-19 Protease Inhibitors Assisted by Molecular...

Area of Science:

• Computational chemistry
• Drug discovery
• Artificial intelligence in medicine

Background:

• The COVID-19 pandemic has overwhelmed global health services.
• Artificial Intelligence (AI) and Machine Learning (ML) are crucial for disease pattern tracking and treatment identification.

Purpose of the Study:

• To identify potential COVID-19 protease inhibitors using shape-based ML, molecular docking, and molecular dynamics simulations.

Main Methods:

• 31 repurposed compounds were selected targeting the main coronavirus protease (6LU7).
• A machine learning approach generated shape-based molecules.
• Ligand-Receptor Docking and Molecular Dynamic Simulations were performed, followed by ADMET studies.

Main Results:

• Remdesivir, valrubicin, aprepitant, and fulvestrant showed high affinity for the target protein.
• A novel compound, nCorv-EMBS, was identified and showed suitable toxicity profiles.
• nCorv-EMBS is a promising candidate for COVID-19 treatment.

Conclusions:

• Inhibitors targeting ACE-II, GAK, AAK1, and protease 3C can block viral entry.
• The novel compound nCorv-EMBS shows promise as a COVID-19 protease inhibitor for further evaluation.