PredHS: a web server for predicting protein-protein interaction hot spots by using structural neighborhood properties

Area of Science:

• Computational biology
• Structural bioinformatics
• Protein-protein interactions

Background:

• Identifying critical residues in protein interfaces is essential for understanding binding affinity and specificity.
• Existing methods for predicting these 'hot spot' residues often have limitations in accuracy and scope.

Purpose of the Study:

• To develop and present PredHS, an interactive web server for predicting protein hot spot residues.
• To offer a highly accurate and effective structure-based prediction method.

Main Methods:

• Integration of novel structural and energetic features.
• Utilizing Euclidean and Voronoi structural neighborhoods.
• Employing random forest and sequential backward elimination for feature selection.

Main Results:

• PredHS demonstrates superior performance in identifying hot spots compared to state-of-the-art methods.
• Benchmarking was conducted using an independent dataset verified by experiments.

Conclusions:

• PredHS provides a powerful and accurate tool for predicting protein hot spot residues.
• The web server enables visualization and download of predictions for protein structures in PDB format.