Determination of topo-geometrical equivalence classes of atoms

In this paper we present a Prolog program which generates a molecular code that identifies the topological and geometrical equivalence classes of atoms in molecules. The initial partitioning of atoms is based on the list of successive extended connectivity and on the chromatism of the molecular graph. The definitive topological classes are obtained by matching all canonical labeled mapping trees exhaustively generated during the construction of a unique one-to-one topological code. Further, the geometric classes of atoms are identified on the basis of the lists of successive interatomic Euclidean distances generated following the molecular topology. The resulting topo-geometrical equivalence classes of atoms allow the construction of a molecular canonical code which maintains the distinction between the topology and the geometry. Thus stereoisomers and conformers share the same topological part of the canonical code but differ by their geometrical parts. The method developed in this work has a wide range of applications in chemical information systems: coding of chemical compounds in three-dimensional databases, searching in molecular databases, generation of substitution and labeled isomers, and the identification of the number of signals and equivalent atoms in NMR and ESR spectra. The algorithm, implemented in a Prolog program, can be used for the automatic recognition of topo-geometrical features of chemical compounds.