Transcriptional data: a new gateway to drug repositioning?

Abstract

Recent advances in computational biology suggest that any perturbation to the transcriptional programme of the cell can be summarised by a proper 'signature': a set of genes combined with a pattern of expression. Therefore, it should be possible to generate proxies of clinicopathological phenotypes and drug effects through signatures acquired via DNA microarray technology. Gene expression signatures have recently been assembled and compared through genome-wide metrics, unveiling unexpected drug-disease and drug-drug 'connections' by matching corresponding signatures. Consequently, novel applications for existing drugs have been predicted and experimentally validated. Here, we describe related methods, case studies and resources while discussing challenges and benefits of exploiting existing repositories of microarray data that could serve as a search space for systematic drug repositioning.

Figure 1

Signature reversion (a) and guilt-by-association (b) approaches in gene-expression-based drug repositioning. In (a) the aim is to identify a drug where the effect on transcription is opposite to a disease signature. In (b) drugs eliciting similar gene expression signatures are sought and hypothesised to share a common mode of action. Many publicly available repositories can be queried to generate drug and disease signatures that can be compared to each other and integrated with newly generated experimental data (c).

Figure 2

Rate of growth of ArrayExpress data in terms of experiments (i.e. user submission). This trend is set to increase further in the future, as new high-throughput sequencing-based transcriptomic applications result in the generation of huge amounts of data.