Enabling the early detection and prevention of diabetic kidney damage has potential to substantially reduce the global burden of kidney failure. There is a critical need for identification of mechanistic biomarkers that can predict progression and serve as therapeutic targets. In this issue of the JCI, Sharma and colleagues used an integrated multiomics approach to identify the metabolite adenine as a noninvasive biomarker of progression in early diabetic kidney disease (DKD). The highest tertile of urine adenine/creatinine ratio (UAdCR) was associated with higher risk for end-stage kidney disease and mortality across independent cohorts, including participants with early DKD without macroalbuminuria. Spatial metabolomics, single-cell transcriptomics, and experimental studies localized adenine to regions of tubular pathology and implicated the mTOR pathway in adenine-mediated tissue fibrosis. Inhibition of endogenous adenine production was protective in a diabetic model. These findings exemplify the potential for multiomics to uncover mechanistic biomarkers and targeted therapies in DKD.