LeadGen is a new system for de novo design based on two stages application of Genetic algorithm to account for molecular flexibility. A chemical-structure generation within ‘knowledge base’ constraints is provided in the first stage, analogously to the Glen and Payne’s (1995) algorithm. The constraints may be any steric and electronic requirements, in terms of parameter ranges, describing, for example, an enzyme active site, pharmacophore or properties from pattern recognition analysis. In a second stage of presented algorithm, GA is employed to investigate conformational flexibility of generated structures, utilizing a recently derived approach for coverage of the conformational space by a limited number of conformers (Mekenyan et al., 1999). The starting point of the system is a random collection of fragment and molecules; known molecules may also be utilized. The evolutionary operators of selection, crossover and mutation are employed to evolve the input structures into new molecules, which better fit the constraints. In the first stage of GA application the chemical ‘consistency’ of molecules is improved, whereas, in the second stage they are conformationally optimized. Thus, one can avoid cases in which a molecule could be neglected just because it is generated in a conformation, which is different than the active one. Glen RC, Payne AWR. 1995. A Genetic Algorithm for the Automated Generation of Molecules within Constraints. J Comput-Aided Mol Des 9:181-202. Mekenyan OG, Dimitrov D, Nikolova N, Karabunarliev S. 1999. Conformational Coverage by a Genetic Algorithm. Chem Inf Comput Sci 39/6: 997-1016.