Analysis of Amino Acid Properties in Interaction Surfaces of Decoys Generated by Re-Docking Scheme

Abstract

Rigid-body docking processes generate many protein complexes (decoys) for searching near-native decoys (NN-decoys) in post-docking processes with analyzing various pairs of surfaces on two input proteins, using properties of electrostatics and desolvations between two molecules or proteins. Many powerful docking computer softwares were developed and were used for resolving various categories of problems, for example, analysis of protein interaction networks or drug design. Rigid-body docking process is popular and useful in such works. However, there are some unfortunate cases, which could not obtain NN-decoys. Then, we developed Re-docking scheme using interaction fingerprints (IFPs). Re-docking scheme is the process of iterating rigid-body docking for generating more NN-decoys. After initial-docking, we classify decoys into several interaction surfaces. Thereafter, other docking processes are performed with more fine searching limited in every interaction surfaces classified. We could obtain NN-decoys even if no NN-decoys in initial-docking process [Uchikoga et. al. (2013) PLOS ONE 8:e69365]. Then, we approach a problem of prediction of protein-protein interactions by using IFPs, which gives us properties of physico-chemical interactions because IFPs are composed of interacting amino acid pairs [Uchikoga & Hirokawa (2010) BMC Bioinform. 11:236]. By using IFPs, we can obtain these properties easily and trace interaction surfaces in docking processes. In this work, docking search spaces become to be seen by using amino acid properties involved in molecular surfaces of many decoys, generating by Re-docking scheme. Then, we would like to discuss about understanding interaction mechanisms.