Drugs, structures, fragments : substructure-based approaches to GPCR drug discovery and design

Abstract

This thesis is all about cheminformatics, and its impact on drug discovery. A number of strategies are discussed that apply computational methods for the analysis and design of G protein-coupled receptor (GPCR) ligands. Frequent substructure mining is applied to find the common structural motifs that are discriminative for predefined classes of GPCR ligands. In addtion, this approach is extended to cluster GPCRs to suggest a new classification for this receptor superfamily. Furthermore, substructure analysis is utilised to screen for new adenosine A2A receptor ligands. Finally, an automated de novo design approach is described that is used for the design of new adenosine A1 receptor ligands using a multi-objective evolutionary algorithm.