Influence of genetic algorithm parameters on the docking efficiency with the SOL program

The key genetic algorithm parameters and their influence on the docking efficiency implemented in the SOL parallel program are analyzed. Tests were carried out with complexes of uPA (urokinase-type plasminogen activator) and co-crystallized ligands and were directed on the research of new inhibitors of uPA as potential antitumor drugs. The optimal parameters of the genetic algorithm are obtained for the single-run mode, i.e., for the docking of a single molecule, and for the virtual screening mode, i.e., for the docking of compound libraries. Tests were performed on the CHEBYSHEV and LOMONOSOV cluster supercomputers.