Scoring functions and their evaluation methods for protein-ligand docking: recent advances and future directions

Area of Science:

• Computational chemistry
• Structural biology
• Drug discovery

Background:

• Scoring functions are vital for structure-based drug design.
• Predicting protein-ligand interactions accurately and rapidly remains a challenge in molecular docking.

Purpose of the Study:

• To review basic types of scoring functions: force-field, empirical, and knowledge-based.
• To discuss the consensus scoring technique in protein-ligand docking.
• To present and discuss assessment criteria and databases for scoring function evaluation.

Main Methods:

• Literature review of scoring function types (force-field, empirical, knowledge-based).
• Analysis of consensus scoring techniques.
• Discussion of assessment criteria and public databases for performance evaluation.

Main Results:

• Overview of three fundamental scoring function categories.
• Explanation of consensus scoring approaches.
• Identification of common evaluation metrics and datasets.

Conclusions:

• Existing scoring functions face challenges in accurate protein-ligand interaction prediction.
• Future research should focus on developing improved scoring functions for molecular docking.