Protein Docking using Constrained Self-adaptive Differential Evolution Algorithm

The objective of protein docking is to achieve a relative orientation and an optimized conformation between two proteins that results in a stable structure with the minimized potential energy. Constrained Self-adaptive Differential Evolution (Cons_SaDE) algorithm is used to find the minimum energy conformation using proposed constraints such as boundary surface complementary interactions, non-bonded inter-atomic allowed distances, and finding of interaction and non-interaction sites. With these constraints, Cons_SaDE is efficient enough to explore the promising solutions by gradually self-adapting the strategies and parameters learnt from their previous experiences. Modified sampling scheme called Rotate Only Representation is used to represent a docking conformation. GROMOS53A6 force field is used to find the potential energy. To test the performance of this algorithm, few bound and unbound complexes from Protein Data Bank (PDB) and few easy, medium and difficult complexes from Zlab benchmark4.0 are used. Buried Surface Area, Root Mean Square Deviation (RMSD) and Correlation Coefficient are some of the metrics applied to evaluate the best docked conformations. RMSD values of the best docked conformations obtained from five popular docking web servers are compared with Cons_SaDE results and nonparametric statistical tests for multiple comparisons with control method are implemented to show the performance of this algorithm. Cons_SaDE has produced good quality solutions for the most of the data sets considered.