Protein-protein docking using multi-dimensional spherical basis functions on high performance computing platform

Abstract

Docking has become the most important in-silico technique in the process of in-silico drug discovery. The complexes produced because of protein-ligand and the protein-protein interaction are predicted using docking techniques. Hence, it is very important but at the same time it is quite challenging owing to the huge computational costs and the complexity of the computational techniques. In our previous work [1], we had studied the role of the FFTs in increasing the efficiency of Scoring Functions (SF) in heterogeneous parallel processing based virtual screening pipeline for effective rescoring in protein-ligand docking. In this work, we propose multi-layered polar transformation functions to search a multi-dimensional space of a rigid-body model. These functions enhance the efficient use of the spherical co-ordinates to improve the scoring function, thereby improving the overall efficiency of the docking process.