JET2 Viewer: a database of predicted multiple, possibly overlapping, protein-protein interaction sites for PDB...

Area of Science:

• Biochemistry
• Structural Biology
• Bioinformatics

Background:

• Protein binding sites are crucial for molecular interactions and biological functions.
• Accurate prediction of these sites is essential for drug discovery and protein engineering.
• Existing methods often lack comprehensive large-scale validation.

Purpose of the Study:

• To introduce the JET2 Viewer, a database of predicted protein binding sites.
• To present the JET² computational method for identifying and analyzing protein-protein interaction surfaces.
• To provide a large-scale evaluation of a protein binding site prediction method.

Main Methods:

• Application of the JET² computational method to over 20,000 Protein Data Bank (PDB) chains.
• Generation of a comprehensive database of putative protein binding sites.
• Evaluation of JET² predictions against more than 15,000 experimentally characterized protein interfaces.

Main Results:

• The JET2 Viewer database provides predictions for all 3D structures in the PDB.
• JET² demonstrates high precision in predicting interacting surfaces and their evolutionary complexity.
• The method achieved an overall accuracy of 75.89% (Sen=52.52, PPV=51.24, Spe=80.05) in predicting protein interfaces.

Conclusions:

• JET2 Viewer offers a valuable resource for researchers studying protein binding sites.
• The JET² method provides precise predictions and insights into the evolutionary aspects of protein interactions.
• The findings can guide strategies for modulating protein-protein interactions and redesigning interaction surfaces.