Automated recognition of common structural patterns among drug molecules.
Yoshimasa Takahashi, Shin‐ichi Sasaki · Journal of Synthetic Organic Chemistry Japan · 1987
It is well known that chemical structure diagram or three-dimensional structure of a compound molecule involves much information related to biological activity, reactivity, or various physicochemical properties on itself. It is a very important thing in a wide range of chemical research to compare the chemical compounds on the basis of such structural information. Especially, in the area of structure-activity studies, to investigate structural feature such as functional groups or substructural fragments, common to a group of chemical compounds exhibiting a particular biological activity often provides us with very useful and objective information which may be resulted in hypothetical pharmacophore.In this article, a couple of approaches for the automated recognition of common structural patterns among drug molecules will be described. One is related to the maximal common substructure search which can be used to find some topological pharmacophore, and the other one is for the common geometrical pattern search in which three-dimensional geometry of molecules are taken account. The latter approach may be used to identify a certain topographical or three-dimensional pharmacophore. The basic algorithms and related software systems will be summarized with some illustrative examples.