Protein-Protein Docking with Improved Shape Complementarity

Protein-protein docking is a useful computational tool for predicting the complex structure and interaction between proteins. As the most basic ingredient of scoring functions, shape complementarity plays a critical role in protein-protein docking. In this study, we have presented a new pairwise scoring function to consider long-range interactions in shape complementarity (LSC) for protein-protein docking. Our docking program with LSC was tested on the protein docking benchmark 4.0 of 176 diverse protein-protein complexes, and compared with four other shape-based docking approaches, ZDOCK2.1, MolFit/G, GRAMM, and FTDock/G. It was shown that our LSC significantly improved the docking performance in binding mode predictions in both success rate and number of hits per complex, compared to the other four approaches. The software is freely available as part of our HDOCK web server at http://hdock.phys.hust.edu.cn/.