Stochastic generator of chemical structure .2. Using simulated annealing to search the space of constitutional isomers

While there have been many theoretical and applied studies to explore the space of conformational isomers, little has been reported regarding constitutional (i.e., structural) isomers. Deterministic algorithms have been proposed in the past but are limited to small molecules because constitutional space sizes scale exponentially with the number of atoms. The present paper proposes a stochastic algorithm based on the simulated annealing method that searches constitutional isomers with desired properties. The algorithm is general enough to be used for any class of organic and inorganic compounds, including cyclic and crosslinked structures. Theoretically, the algorithm is shown to be efficient (i.e., polynomial). Practically, the algorithm performs remarkably well even in very large constitutional space sizes (up to 10(32)). Applications of this algorithm are suggested in the context of computer-aided molecular design. The practical examples given in the paper include the search for structures having specific topological indices, the search for structures having low log P (octanol/water) partition coefficient, and the search for global minimum of energy in constitutional spaces. The proposed algorithm appears to be the first one that searches chemical structures without restricting the search space.