GANM: A protein–ligand docking approach based on genetic algorithm and normal modes

We describe a new method and a software, called GANM (Genetic Algorithm with Normal Modes), that combines genetic algorithm and normal mode analysis in order to achieve protein–ligand docking simulations that take into account global and local conformational adjustments. The energetical and structural optimization during the ligand binding process is achieved with GA that uses a rotamer library for changing the side chain conformations within and around the binding site. The methodology uses normal modes to account for the global conformational changes of the protein. This software is implemented in distinct versions using different levels of approximations. In this article, we describe the methodology and discuss the results and the performance obtained with the first two versions of GANM. The protein–ligand complexes considered for the test calculations are: (1) Dihydrofolate Reductase (DHFR) and MTX as ligands, (2) HIV-1 Aspartyl Protease with ritonavir, nelfinavir and DMP323 as ligands, and (3) CDK-2 and DT-5 as ligands.