Predicting FFAR4 agonists using structure-based machine learning approach based on molecular fingerprints.

Area of Science:

• Pharmacology
• Computational Chemistry
• Biophysics

Background:

• Free Fatty Acid Receptor 4 (FFAR4) is a G-protein-coupled receptor involved in regulating physiological processes.
• FFAR4 agonists may enhance insulin release and reduce metabolic disease risks.

Purpose of the Study:

• To identify novel FFAR4 agonists using molecular structure-based machine learning.
• To validate potential agonists through molecular docking, ADME/Toxicity predictions, and molecular dynamics simulations.

Main Methods:

• Machine learning (Bayesian network) applied to molecular fingerprints for initial screening.
• Molecular docking and ADME/Toxicity predictions for hit validation.
• 100 ns Molecular Dynamics (MD) simulations of FFAR4-ligand complexes.

Main Results:

• Machine learning identified promising candidate compounds.
• MD simulations revealed stable FFAR4-ligand complexes with significant interactions at crucial residues.
• Analyses indicated compact structures (RMSD 3.57-3.64 nm, fluctuations 5.27-6.03 nm) and thermodynamic stability.

Conclusions:

• Two compounds, CHEMBL2012662 and CHEMBL64616, are identified as potential FFAR4 agonists.
• These compounds warrant further investigation for therapeutic applications in metabolic and immune disorders.