A fuzzy-set approach to functional-group comparisons based on an asymmetric similarity measure

A new, fuzzy-set framework for molecular similarity analysis is proposed. R ecent advances in the computation of ab initio quality fuzzy electron densi ties of molecular fragments, pharmacophoric moieties, and functional groups of both small and macromolecular systems have motivated the search for new approaches to the evaluation of molecular similarity. Fuzzy-set methods ap pear especially suitable for the task for two reasons: the inherent fuzzine ss of electron densities due to Heisenberg's quantum mechanical uncertainty relation, and the fuzzy range of possible local molecular conformations wh ich may correlate with a given chemical or biochemical function. Ln this co ntribution, a new, asymmetric similarity measure is proposed for local mole cular moieties, such as functional groups or pharmacophores, based on a fuz zy-set description of electron densities and a fuzzy-set adaptation of earl ier scaling-nesting similarity measures. (C) 1999 John Wiley & Sons, Inc.