Figure 1. The project analysis process.

The three major phases of the project are defined by the grey boxes. A limited number of gene signatures were considered (four) to limit false-positive associations. The compound (CMPD) selection phase first confirmed that the drug repurposing signature provided valid matches with in vivo positive control drugs, and then a full list of in vitro active drug matches was generated. The third phase was an iterative process in that validation was considered on several levels. We utilised four main independent validation strategies, incorporating multiple data sources, to demonstrate that the insulin resistance drug repurposing (IR-DR) signature produced a high rate of likely true-positive drugs that would reverse IR.