LGA-Dock/EM-Dock Exploring Lamarckian genetic algorithms and energy-based local search for ligandreceptor docking

We report the development of LGA-Dock and EM-Dock, two SVL-based docking programs for flexible ligand – rigid receptor docking applications. LGA-Dock is comprised of a stochastic population generator, a docking routine based on a Lamarckian genetic algorithm, and a local search function based on molecular mechanics (MM) energy minimization. Subsequent modifications of LGA-Dock to address performance issues produced EM-Dock, which proved to be as accurate and much faster than its predecessor despite the deletion of the genetic algorithm component. The basic performance of LGA-Dock and EM-Dock, compared with AutoDock and MOE™ 2004.03 docking routines is presented.Key words: docking, Lamarckian genetic algorithm, molecular mechanics, simulated annealing, tabu search, local search.