Search for approaches to improving the calculation accuracy of the protein—ligand binding energy by docking

Abstract

The following reasons limiting the accuracy of calculations of the protein—ligand binding energy by the molecular docking programs are considered: the limited facilities of algorithms of finding a global minimum on a complicated multi-dimensional protein—ligand energy surface, restrictions on the degrees of freedom of a protein—ligand system including docking into a rigid protein, inadequacy of the existing force fields, a lack of taking into account the solvent or too rough allowance for the solvent in the docking procedure, a lack of the local energy optimization in the docking process, an inaccuracy of the construction of models of a target protein and a ligand, simplification of the calculation method of the Gibbs free energy of a molecular system, and limited computer resources for docking of one ligand. A new approach to the development of the new generation of docking programs is proposed. The approach allows one to remove step-by-step the existing simplifications and to increase considerably the accuracy of the whole docking process, including the calculation accuracy of the protein—ligand binding energy. The results of the study are presented and demonstrate the computational feasibility of the assigned docking problem.