Improvement of the protein-protein docking prediction by introducing a simple hydrophobic interaction model: An application to interaction pathway analysis

Abstract

We propose a new hydrophobic interaction model that applies atomic contact energy for our protein-protein docking software, MEGADOCK. Previously, this software used only two score terms, shape complementarity and electrostatic interaction. We develop a modified score function incorporating the hydrophobic interaction effect. Using the proposed score function, MEGADOCK can calculate three physico-chemical effects with only one correlation function. We evaluate the proposed system against three other protein-protein docking score models, and we confirm that our method displays better performance than the original MEGADOCK system and is faster than both ZDOCK systems. Thus, we successfully improve accuracy without loosing speed. © 2012 Springer-Verlag.