Recent technical developments have enabled large-scale molecular analyses of biological samples, such as the assessment of plasma metabolomics and proteomics. Some of these compounds have been found to be either increased or decreased in individuals with diabetes and pre-diabetes, but it is unclear whether these aberrations are part of the pathogenesis. We and others have applied bi-directional Mendelian Randomization methods in population-based cohorts and public genetic association data of metabolomics and glycemic traits to try to disentangle the causal directions. Given the vast effects of insulin on the metabolism, it is not surprising to find genetic evidence that several of these aberrations are secondary to insulin resistance. In many cases, however, especially for lipid metabolites, it has been difficult to find specific genetic instruments to assess causality.