ESSESA - AN EXPERT-SYSTEM FOR STRUCTURE ELUCIDATION FROM SPECTRA .4. CANONICAL REPRESENTATION OF STRUCTURES

LNSCS (linear notation system of chemical structures), a chemical structure representation system, is designed to support the arbitrary notation for structures in ESSESA (expert system for structure elucidation from spectra). This paper describes a method for the canonical representation of chemical structures, the canonical renumbering of structures, the unique notation and the canonicalization algorithm, and the conversion from LNSCS to the canonical connection table. The first stage of the algorithm is the partitioning of the atoms in a structure, the molecule being treated as a graph with nodes (atoms) and edges (bonds). The partition is done by a product of the extended order value method; the initial order value is generated from the atomic character of the atoms in the structure. The program of converting LNSCS code to the canonical connection table consists of three parts, the compiling procedure of LNSCS and partitioning and generation of the canonical connection table. The algorithm by product of extended order value method is found to be more effective than the Morgan algorithm and other published algorithms.