Selection-Insertion Schemes in Genetic Algorithms for the Flexible Ligand Docking Problem

Camila S. de Magalhães, Helio J.C. Barbosa, and Laurent E. Dardenne

Abstract. In this work we have implemented and analyzed the performance of a new real coded steady-state genetic algorithm (SSGA) for the flexible ligand-receptor docking problem. The algorithm employs a grid-based methodology, considering the receptor rigid, and the GROMOS classical molecular force field to evaluate the energy function. In the implementation we used the restricted tournament selection (RTS) technique in order to find multiple solutions and also introduced a new variation of this technique with an insertion criterion based in the root-mean-square-deviation (RMSD) between the ligand structures. The SSGA was tested in docking four HIV1 protease-ligand complexes with known three-dimensional structures. All ligands tested are highly flexible, having 12 to 20 conformational degrees of freedom. The implemented docking methodology was able to dock successfully all flexible ligands tested with a success ratio higher than 90 % and a mean RMSD lower than 1.3 Å with respect to the corresponding experimental structures.

LNCS 3102, p. 368 ff.

Full article in PDF

lncs@springer.de
© Springer-Verlag Berlin Heidelberg 2004