For the study, appearing in the journal Science Translational Medicine, the team constructed a signature for Type 2 diabetes based on 50 genes, then used publically available expression datasets to screen 3,852 compounds for drugs that potentially reverse disease.