A comprehensive survey on protein-ligand binding site prediction

Area of Science:

• Computational biology
• Drug discovery
• Structural bioinformatics

Background:

• Protein-ligand binding site prediction is crucial for understanding biological processes and identifying drug candidates.
• Experimental methods are resource-intensive, necessitating efficient computational approaches.
• Accurate prediction aids in protein function annotation and accelerates the drug development pipeline.

Purpose of the Study:

• To review current challenges in ligand binding site (LBS) prediction.
• To summarize recent computational methods, analyzing input features, algorithms, and ligand types.
• To explore allosteric site identification as a specific LBS type and discuss future machine learning trends.

Main Methods:

• Literature review of recently published computational methods for LBS prediction.
• Analysis of method inputs, algorithms, and applicable ligand classes.
• Focused investigation on allosteric site prediction techniques.

Main Results:

• Identified key challenges in LBS prediction, including accuracy and efficiency.
• Categorized and compared various computational approaches based on their features and algorithms.
• Highlighted the importance and specific considerations for allosteric site identification.

Conclusions:

• Computational methods are vital for overcoming limitations of experimental LBS prediction.
• Machine learning holds significant promise for advancing LBS prediction accuracy and scope.
• Future research should focus on refining algorithms and expanding the application of LBS prediction in drug discovery.