TOPOLOGICAL STRUCTURAL INFORMATION IN THE CAS FILE - STATISTICAL OCCURRENCES OF DARC CONCENTRIC FRAGMENTS .1. BASIC CARBON SUBSTRUCTURES

The Chemical Abstracts Service (CAS) Chemical File of some 7.5 million structures, structured in DARC topological space, was analyzed statistically for the frequency of occurrence of different topological substructures. These topological DARC fragments, known as FREL-Bs, are described concentrically around a single carbon atom with an environment limited to the first (A) and second (B) neighbor atoms and treated as a spanning graph tree. Occurrence searches were carried out for the 70 primitive carbon FRELs based on σ bonding. These represent potential acyclic or cyclic substructures with an ordered environment and are called FREL-AC or FREL-CY, respectively. Such ordered environments provide more than mere statistics of local and global descriptions; they lead to trend analyses. Certain broad correlations, linked to the idea of structural neighbors, can thus be observed. Analysis of the statistics associated with cyclic structures revealed some counterintuitive exceptions to the general rule that occurrence should be inversely related to complexity. As an example, the presence of various familiar chemical families such as steroids and terpenes can be detected within the context of the file and can be tagged by means of specific FREL associations. © 1990, American Chemical Society. All rights reserved.