Open Binding Pose Metadynamics: An Effective Approach for the Ranking of Protein-Ligand Binding Poses

Area of Science:

• Computational chemistry
• Molecular modeling
• Drug discovery

Background:

• Accurate prediction of ligand-protein binding poses and affinities is crucial for computer-aided drug discovery.
• Existing methods like molecular docking and enhanced sampling simulations have limitations in speed and accuracy.
• Binding Pose Metadynamics (BPMD) offers a computationally efficient approach to probe binding affinity by analyzing ligand resistance to simulation bias.

Purpose of the Study:

• To introduce OpenBPMD, an open-source Python reimplementation of the Binding Pose Metadynamics (BPMD) method.
• To enhance the computational efficiency and accuracy of predicting ligand-protein binding interactions.
• To validate the performance of OpenBPMD against the original BPMD and explore improvements through water modeling.

Main Methods:

• Developed OpenBPMD, an open-source Python tool utilizing the OpenMM simulation engine.
• Implemented a revised scoring function within the BPMD framework.
• Validated the algorithm on diverse targets and investigated the impact of grand-canonical Monte Carlo (GCMC) for water positioning.

Main Results:

• OpenBPMD demonstrates performance matching or exceeding the original BPMD in predicting binding poses and stability.
• The study confirmed the utility of BPMD for reranking docking poses.
• Incorporating accurate water positioning, particularly with GCMC, significantly improves prediction accuracy.

Conclusions:

• OpenBPMD provides a valuable, open-source resource for accelerating drug discovery through improved ligand-protein binding predictions.
• The revised scoring function and integration with advanced water modeling enhance the reliability of binding affinity estimations.
• This work facilitates more efficient and accurate virtual screening and lead optimization in pharmaceutical research.