Diversity Profiling and Design Using 3D Pharmacophores:  Pharmacophore-Derived Queries (PDQ)

Abstract

The current interest in combinatorial chemistry for lead generation has necessitated the development of methods for design and evaluation of the diversity of the resultant compound libraries. Such methods also have application in selecting diverse sets of compounds for general screening from corporate databases and in the analysis of large sets of structures to identify common patterns. In this paper we describe a novel methodology for calculating diversity and identifying common features based on the three-point pharmacophores expressed by a compound.1 The method has been implemented within the environment of the Chem-X molecular modeling package (ChemDBS-3D), using a systematic analysis of 3D distance space with three point combinations of six pharmacophoric groups. The strategy used to define the pharmacophores is discussed, including an in-house developed atom type parameterization. The method is compared with the related approach being developed into the ChemDiverse module of Chem-X. Results from an analysis of a large corporate database and examples of combinatorial library profiling with both methods are presented. The use of 3D pharmacophores for assessing diversity, and the application of such methods to combinatorial library design, is discussed.