Predicting Binding Poses and Affinities in the CSAR 2013-2014 Docking Exercises Using the Knowledge-Based Convex-PL...

Area of Science:

• Computational chemistry and structural biology
• Molecular modeling and drug discovery

Background:

• Accurate prediction of protein-ligand interactions is crucial for drug discovery.
• Knowledge-based potentials leverage structural data for improved modeling.

Purpose of the Study:

• To evaluate the performance of a new knowledge-based potential, Convex-PL.
• To assess its efficacy in protein-ligand docking pose prediction and binding affinity ranking.

Main Methods:

• Developed Convex-PL using structural data from the PDBBind database.
• Tested Convex-PL in the 2013-2014 Comparative Assessment of Scoring Functions (CSAR) docking exercise.
• Evaluated performance in near-native pose detection and binding affinity prediction.

Main Results:

• Convex-PL demonstrated high accuracy in near-native pose detection, correctly predicting poses in both 2013 and 2014 CSAR exercises.
• Achieved a Spearman correlation coefficient > 0.5 for several protein-ligand sets in affinity ranking, indicating fair performance.
• Binding affinity prediction remains a challenge, necessitating further methodological improvements.

Conclusions:

• The Convex-PL potential shows significant promise for accurate protein-ligand pose prediction.
• Further refinement is required to enhance its capabilities in predicting binding constants.
• The study highlights the potential of structural data-driven approaches in molecular modeling.