Information theory-based scoring function for the structure-based prediction of protein-ligand binding affinity.

Area of Science:

• Computational Chemistry
• Structural Biology
• Drug Discovery

Background:

• Accurate prediction of binding energy is crucial for identifying potential drug candidates.
• Existing scoring functions often struggle with precise binding energy estimations.
• Knowledge-based scoring functions leverage experimental data for improved accuracy.

Purpose of the Study:

• To develop and validate a novel knowledge-based scoring function, SIScoreJE, for predicting protein-ligand binding energy.
• To assess the impact of solvation effects on binding energy predictions.
• To compare the performance of SIScoreJE against established scoring functions.

Main Methods:

• Derived protein-ligand atomic contact information from a Non-Redundant Data set (NRD) of over 3000 X-ray crystal structures.
• Calculated atomic contact preferences using joint entropy and evaluated 18 atom-type classification schemes, selecting 'ScoreJE Atom Type set2' (SATs2).
• Incorporated Single-body Solvation Potentials (SSP) derived from protein-water interactions and validated using a 100-complex dataset.

Main Results:

• The combined SIScoreJE (SSP/ScoreJE) demonstrated significantly improved performance over ScoreJE alone.
• SIScoreJE and ScoreJE outperformed GOLD::GoldScore, GOLD::ChemScore, and XScore in predicting binding affinities.
• 'ScoreJE Atom Type set2' (SATs2) was identified as the most suitable atom-type classification scheme.

Conclusions:

• SIScoreJE provides an efficient and accurate method for predicting protein-ligand binding energy.
• The inclusion of solvation potentials substantially enhances prediction accuracy.
• SIScoreJE represents a valuable advancement for computational drug discovery and molecular modeling.