AI-Driven Discovery of Prototype CLEC4M Inhibitors Targeting Marburg Virus Entry via Integrated Machine Learning and...

Area of Science:

• Virology
• Computational Chemistry
• Drug Discovery

Background:

• Marburg virus (MARV) causes severe hemorrhagic fever with high mortality.
• Effective therapeutics for MARV are currently unavailable.
• MARV glycoprotein (GP) interaction with host receptor CLEC4M (L-SIGN) is a key target for intervention.

Purpose of the Study:

• To identify novel phytochemical inhibitors targeting the MARV-CLEC4M interaction.
• To utilize machine learning and molecular modeling for drug discovery.

Main Methods:

• Machine learning models (Random Forest) were trained and validated using chemical databases.
• Virtual screening of 11,032 phytochemicals was performed.
• Molecular docking, ADMET analysis, DFT, and MD simulations were employed to evaluate lead compounds.

Main Results:

• A Random Forest model achieved high performance (AUC=0.93, MCC=0.88).
• Virtual screening identified 120 potential active compounds, with 42 meeting drug-likeness criteria.
• Three lead compounds (PubChem IDs: 42608095, 5281601, 11243993) showed promising binding affinities and favorable ADMET profiles.

Conclusions:

• The study successfully identified potential phytochemical inhibitors for MARV entry.
• Combining computational methods is effective for discovering antiviral agents.
• Further experimental validation is needed to develop new MARV therapeutics.