Project Summary/Abstract It is estimated that ketosis-prone diabetes mellitus (KPDM) affects 20-50% of African American patients with newly diagnosed diabetes who present with diabetic ketoacidosis (DKA). At presentation of DKA, these patients have a severe decompensation in insulin secretion accompanied by severe insulin resistance. Unlike patients with type 1 diabetes, following intensive insulin treatment, many of these patients exhibit improvements in insulin secretion and insulin sensitivity and are able to discontinue insulin therapy (near- normoglycemia remission, HbA1c < 7%, fasting glucose < 130 mg/dl while off insulin for at least one week). The period of near-normoglycemia remission is variable and many patients eventually experience a hyperglycemic relapse or even DKA while some stay in remission due to sustained insulin secretion. Despite the phenotype being recognized for many years, the underlying mechanisms leading to ketosis and maintenance of beta-cell function are unknown. Further, increased intra-organ (pancreas and liver) fat can also play a role in remission. High resolution metabolomics and magnetic resonance imaging offer a comprehensive way to assess multiple pathways and imaging characteristics. This proposal will leverage already collected samples and participants from the PI’s K23 proposal to examine pathways leading to ketosis and sustained remission. Aim 1 will compare metabolomic signatures between people with DKA and hyperglycemia without ketosis. Aim 2 will compare metabolomic signatures at time of insulin discontinuation and their relationship with sustained beta-cell function. Aim 3 will explore whether differences in pancreatic and hepatic fat associate with sustained remission and beta-cell function. The Aims, together with new collaborators with complementary expertise will generate data to explore novel hypotheses for new independent proposals by the PI.