Inclusion of the orientational entropic effect and low-resolution experimental information for protein-protein docking in CAPRI

Inclusion of entropy is important and challenging for protein-protein binding prediction. Here, we present a statistical mechanics-based approach to empirically consider the effect of orientational entropy. Specifically, we globally sample the possible binding orientations based on a simple shape-complementarity scoring function using an FFT-type docking method. Then, for each generated orientation we calculate the probability through the partition function of the ensemble of accessible states, which are assumed to be represented by the set of nearby binding modes. For each mode, the interaction energy is calculated from our ITScorePP scoring function that was developed in our laboratory based on principles of statistical mechanics. Using the above protocol, we present the results of our participation in Rounds 22–27 of the CAPRI (Critical Assessment of PRedicted Interactions) experiment for ten targets (T46-T58). Additional experimental information, such as low-resolution SAXS data, was used when available. In the prediction (or docking) experiments of the ten target complexes, we achieved correct binding modes for six targets: one with high accuracy (T47), two with medium accuracy (T48 and T57), and three with acceptable accuracy (T49, T50, and T58). In the scoring experiments of seven target complexes, we obtained correct binding modes for six targets: one with high accuracy (T47), two with medium accuracy (T49 and T50), and three with acceptable accuracy (T46, T51, and T53).