Protein-ligand docking based on β-shape

Abstract

Protein-ligand docking is to predict a ligand conformation and orientation with respect to a protein within its binding site, and has been known to be essential for the development of new drugs. The protein-ligand docking problem is usually formulated as an energy minimization problem to identify the docked conformation of the ligand. A ligand usually docks around a depressed region, called a pocket, on the surface of protein. Presented in this paper is a docking algorithm based on the newly developed geometric construct called β-shape. To cope with the computational intractability, the global optimum is searched using the genetic algorithm. The proposed algorithm first locates initial chromosomes at some locations within the pocket recognized according to the local shape of β-shape. Then, the algorithm proceeds generations by taking advantage of powerful properties of β-shape to achieve a extremely fast and good solution. We claim that the proposed method is much faster than other popular docking softwares including AutoDock.