Type 2 diabetes has a heterogeneous pathophysiology, and is treated with various glucose-lowering medications from classes that differ in their mechanisms of action. Variability in treatment efficacy may be due to genetic variation, but data are needed to guide specific pharmacotherapy addressing the pathogenic mechanisms affecting an individual patient. The Glycemia Reduction Approaches in Diabetes: A Comparative Effectiveness (GRADE) Study is evaluating the effectiveness of each of the four most commonly employed glucose-lowering medications (sulfonylurea, GLP-1 receptor agonist, DPP-4 inhibitor, basal insulin) added to a background of metformin. Among 5,047 randomized participants, 4,730 have consented to genetic analyses. We propose to use a genome-wide association approach to test 1) whether specific genomic regions or polygenic scores associate with the glycemic response to each of the four drugs, or to a specific mode of action; 2) whether specific genomic regions or polygenic scores associate with key intermediate traits of glucose homeostasis (e.g. insulin sensitivity, β- cell function), or the occurrence of side effects and diabetes complications; 3) in a subset of 1,600 participants with available DNA who have undergone continuous glucose monitoring, whether specific genomic regions or polygenic scores associate with the mismatch between glycated hemoglobin and average glycemia, or with glycemic variability while on each drug; and 4) whether variants known to be associated with type 2 diabetes or related traits, polygenic scores constructed from such variants, or physiology-driven partitioned genetic scores, are associated with response to each of the four agents employed in GRADE. This proposal represents a full pharmacogenetic exploration in the GRADE clinical trial designed to advance precision medicine in type 2 diabetes.