Conformer clustering algorithm and its application for crown-type macrocycles

An efficient conformer clustering algorithm designed for treating molecular modeling problems of flexible molecules associated with a large variety of conformers is described. It incorporates a fast algorithm for the generation of chemical graph automorphisms allowing to easily discard equivalent conformers for symmetrical structures. The cluster is defined as a set of conformers deviating from the "central conformer" less than the allowed threshold. The deviation is calculated as a maximum distance between the matched atoms. Matched atoms can include all heavy atoms that can be superimposed according to graph automorphism mappings or a special subset of atoms indicated by the user. Using the "central conformer" as a meaningful representative of a cluster significantly reduces molecular modeling efforts without losing the features of the system. Application of the algorithm to the ion recognition study of the newly synthesized macrocyclic host is presented.