Similarity Measures for Rational Set Selection and Analysis of Combinatorial Libraries: The Diverse Property-Derived (DPD) Approach

Abstract

The generation of new chemical leads for biological targets is a very challenging task for researchers in the pharmaceutical industry. The design of representative screening sets and combinatorial libraries is central to achieving this objective. In this paper, we describe a novel molecular descriptor, the Diverse Property-Derived (DPD) code, that contains information about key molecular and physicochemical properties of a molecule. The utility of this descriptor is explored through its application for the selection of a maximally diverse representative screening set, through the selection of secondary screening sets to obtain more information concerning the structure-activity relationships (SAR) of a particular target receptor, and through the profiling of combinatorial libraries. The usefulness of physicochemical/molecular property descriptors, such as the DPD code, is discussed critically.