An all atom energy based computational protocol for predicting binding affinities of protein-ligand complexes

Area of Science:

• Computational chemistry
• Structural biology
• Drug discovery

Background:

• Accurate prediction of protein-ligand binding affinities is crucial for drug discovery.
• Existing methods can be computationally expensive or lack transferability.

Purpose of the Study:

• To develop a computationally fast and transferable protocol for predicting binding affinities of non-metallo protein-ligand complexes.
• To validate the protocol's predictive accuracy across diverse systems.

Main Methods:

• An all-atom energy-based empirical scoring function was developed.
• The function incorporates electrostatics, van der Waals, hydrophobicity, and loss of conformational entropy.
• The method was trained and validated on a heterogeneous dataset of 161 protein-ligand complexes.

Main Results:

• A high correlation (r=0.92) was achieved between predicted and experimental binding affinities on the training set (61 complexes).
• The model maintained high predictive accuracy (r=0.92) when tested on an independent set of 100 complexes.
• The scoring function demonstrated transferability across diverse protein targets.

Conclusions:

• The developed protocol provides a computationally efficient and accurate method for predicting protein-ligand binding affinities.
• The methodology shows significant potential for application in drug design endeavors.
• The scoring function is available as a web server (BAPPL).