COACH-D

COACH-D predicts protein-ligand binding sites and refines ligand poses using molecular docking to support protein function annotation and drug discovery.

Key Features:

• Binding Site Prediction: Utilizes five distinct methods from the COACH algorithm to predict ligand-binding pockets and residues on protein surfaces.
• Molecular Docking Integration: Incorporates AutoDock Vina to dock and refine ligand poses, producing physically realistic complexes and reducing steric clashes.
• Input Options: Accepts amino acid sequences or three-dimensional protein structures and supports user-submitted ligands or predefined ligand templates for docking into predicted pockets.
• Performance Validation: Validated by benchmark tests and blind evaluations and reports an 85% reduction in steric clashes between ligands and protein structures after refinement.

Scientific Applications:

• Protein Function Annotation: Identifies ligand-binding pockets and residues to support annotation and interpretation of protein function.
• Drug Discovery and Design: Identifies potential drug targets and refines ligand poses to aid design of molecules with improved binding characteristics.

Methodology:

Initial prediction uses the COACH algorithm (five distinct methods) to identify probable binding sites, followed by molecular docking with AutoDock Vina to position user-submitted or template ligands into predicted pockets and refine ligand poses, reducing steric clashes.