Blind docking method combining search of low-resolution binding sites with ligand pose refinement by molecular...

Area of Science:

• Computational chemistry
• Structural biology
• Drug discovery

Background:

• Accurate prediction of protein-ligand binding poses is crucial for rational drug design.
• Existing molecular docking methods face challenges in accurately predicting binding poses, especially in blind docking scenarios.

Purpose of the Study:

• To develop and validate a novel blind hierarchical docking method, bhDock, for improved protein-ligand interaction prediction.
• To assess the accuracy and performance of bhDock compared to established docking algorithms.

Main Methods:

• bhDock employs a two-step hierarchical approach: low-resolution binding site identification and global optimization using molecular dynamics.
• Ligand pose refinement involves simulated annealing molecular dynamics and guided force-field deformation for global minimum identification.

Main Results:

• The bhDock method demonstrated a 78% success rate in predicting correct binding poses for 37 protein-ligand complexes.
• This success rate surpasses that of widely used docking tools like AutoDock, DOCK, GOLD, and FlexX on the same dataset.

Conclusions:

• bhDock represents a significant advancement in blind hierarchical docking, offering enhanced accuracy for protein-ligand binding pose prediction.
• The method's performance suggests its potential utility in accelerating drug discovery and development pipelines.