Type 2 diabetes mellitus affects 9.6% of the adults in the United States and more than 200 million people worldwide. Diabetes can be a devastating disease, but it can now be treated with nine classes of approved drugs (insulins, sulfonylureas, glinides, biguanides, alpha-glucosidase inhibitors, thiazolidinediones, glucagon-like peptide 1 mimetics, amylin mimetics, and dipeptidyl peptidase 4 inhibitors), in addition to diet and exercise regimens. Choosing which drug to give a patient is based on efficacy and also availability, cost, safety, tolerability, and convenience. Personalized medicine promises a path for individually optimized treatment choices, but realizing this promise will require a more comprehensive characterization of disease and drug response. In this issue of the JCI, Shu et al. make significant progress by integrating diverse data supporting the hypothesis that genetic variation in organic cation transporter 1 (OCT1) affects the response to the widely used biguanide metformin (see the related article beginning on page 1422). We discuss metformin, OCT1, pharmacogenetics, and how the integrative genomics revolution is likely to change our understanding and treatment of diabetes.