DIVSEL and COMPLIB--strategies for the design and comparison of combinatorial libraries using pharmacophoric descriptors

Abstract

Screening synthetic combinatorial libraries may facilitate rapid drug lead discovery by substantially increasing the number of molecules tested. Drug discovery efficiency and productivity can be further improved by designing libraries to maximize their molecular diversity or by comparing them to existing collections of compounds and/or libraries to select those that complement the properties already well represented. In this paper we describe two strategies to aid in the design and comparison of combinatorial libraries. The methods employ multi-pharmacophore three-dimensional (3D) descriptors in combination with two recent proposals for dissimilarity-based compound selection and library comparison. This method allows the design to be performed in product space and library comparison to consider all pair-wise intermolecular contributions to the diversity.