HarmonyDOCK: the structural analysis of poses in protein-ligand docking

Area of Science:

• Computational chemistry
• Structural biology
• Drug discovery

Background:

• Molecular docking is essential for lead optimization in drug discovery.
• Current docking tools struggle with accurate prediction of protein-ligand binding poses.
• Reliable pose prediction is critical for identifying effective drug candidates.

Purpose of the Study:

• To introduce HarmonyDOCK, a novel method for evaluating molecular docking software accuracy.
• To develop a consensus scoring function for determining the most accurate protein-ligand pose from multiple docking programs.
• To enhance the reliability of computational methods in drug design.

Main Methods:

• A benchmark set of protein-ligand complexes with known 3D structures was utilized.
• Multiple docking programs were employed to predict ligand conformations.
• HarmonyDOCK was developed to rank poses and establish scoring function parameters.
• Focus was placed on analyzing top-ranked poses for accuracy.

Main Results:

• Significant variations in performance were observed among different docking programs.
• HarmonyDOCK demonstrated the necessity of proper decoy selection for successful docking.
• The method successfully identified a consensus pose among predictions from various docking tools.
• A reference point and lower score limit were established for pose assessment.

Conclusions:

• HarmonyDOCK offers a robust approach to assess docking software accuracy and improve protein-ligand binding predictions.
• The consensus scoring function refines pose selection, enhancing the reliability of docking studies.
• This method is vital for advancing lead optimization and accelerating drug discovery pipelines.