Integrating atom-based and residue-based scoring functions for protein-protein docking.

Area of Science:

• Computational Biology
• Structural Bioinformatics
• Biochemistry

Background:

• Protein-protein docking relies on scoring functions to evaluate binding poses.
• Existing atom-based functions offer high accuracy, while residue-based functions accommodate conformational flexibility.
• Previous work developed ZRANK, an atom-based scoring function for docking prediction reranking.

Purpose of the Study:

• To develop an improved scoring function for protein-protein docking by integrating diverse potential types.
• To enhance the accuracy and success rate of protein complex structure prediction.

Main Methods:

• Combined ZRANK's atom-based potentials with five residue-based potentials and the IFACE atom-based potential.
• Optimized weights for scoring function terms using decoys from the ZDOCK algorithm.
• Validated scoring function combinations using 96 test cases from a protein-protein docking benchmark.

Main Results:

• The integrated scoring function, IRAD, demonstrated improved performance over ZDOCK and ZRANK.
• Addition of IFACE and a residue-based potential reduced cases without correct predictions by 38% and 27%, respectively.
• The IRAD scoring function significantly enhances the integrative success rate for protein-protein docking predictions.

Conclusions:

• Integrating residue-based and atom-based potentials into a single scoring function improves protein-protein docking performance.
• The developed IRAD scoring function offers a more effective approach for predicting protein complex structures.
• IRAD is available for use in protein-protein docking studies.