Refining and implementing the genomic analysis technique of connectivity mapping

Microarray analysis measures the expression of thousands of mRNAs simultaneously, thereby creating genome-wide transcriptional snapshots of biological states. The novel technique of connectivity mapping is a tool that identifies functional chemical similarities by identifying similarities in the transcriptional signatures they induce. As this is a relatively new technique, the aims of this work were two-fold: Firstly, to optimise the method of connectivity mapping, and secondly to establish its potential uses, by applying the technique in several model systems. Connectivity mapping currently requires query signatures to be in Affymetrix geneidentifier format, thus, gene-expression profiles from other microarray platforms must be converted to Affymetrix identifiers, which can introduce error. Converting the connectivity map database into a universal gene-naming nomenclature reduces the error and allows expansion of the database, making use of existing microarray data resources. The use of gene-expression patterns for disease states, combined with connectivity mapping, to identify potential therapeutic agents, was explored. Small molecules capable of sensitising doxorubicin-resistant cells to doxorubicin and preventing cell death in several Huntington’s disease models were discovered. This provided new insights into therapeutic targets for these conditions and suggests the exciting possibility that transcription patterns can be used to establish and identify novel drugtarget relationships. To discover functional chemical similarities, connectivity mapping was used to investigate the effects of the small molecule apogossypol, which to induces endoplasmic reticulum (ER) aggregation. Connectivity mapping facilitated the discovery of a diverse subset of chemicals that also induced ER aggregation, some of which had been extensively functionally characterised. This provided both mechanistic insight and physiological relevance to the ER aggregation phenomenon. This work clearly demonstrates that, although in its infancy, connectivity mapping is an invaluable resource with potential applications at many stages of the drug discovery process, from cellular pathway dissection and drug target identification, to drug discovery and toxicological assessment