New strategies to predict protein-peptide interactions

Peptides are short polymer chains consisting of amino acids. They are flexible and often change conformations when they bind to proteins. Protein-peptide interactions play an important role in many cellular processes. In silico prediction of protein-peptide complex structure is highly desirable for mechanistic investigation of these processes and for therapeutic design. However, it is challenging to predict all-atom structures of protein-peptide complexes without any knowledge about the binding site and the bound peptide conformation, because of the large degrees of freedom involved in the system. In this talk, I will present our recent development of new strategies for predicting protein-peptide complex structures, based on the integration of information-driven modeling and physics/chemistry-based computational modeling of the interaction modes. The peptide is treated as a flexible structure during modeling, and the search space includes the whole surface of the protein. The methods have been systematically and extensively tested, and the results will be presented. The methods are computationally efficient. They can be used either as a standing-alone tool for large-scale protein-peptide docking or as an initial-stage sampling tool for protein-peptide structure refinement programs.