The pace of deorphanization of G protein-coupled receptors (GPCRs) has slowed, and new approaches are required. Small molecule targeting of orphan GPCRs can potentially be of clinical benefit even if the endogenous receptor ligand has not been identified. Many GPCRs lack common variants that lead to reproducible genome-wide disease associations, and rare-variant approaches have emerged as a viable alternative to identify disease associations for such genes. Therefore, our goal was to prioritize orphan GPCRs by determining their associations with human diseases in a large clinical population. We used sequence kernel association tests to assess the disease associations of 85 orphan or understudied GPCRs in an unselected cohort of 51,289 individuals. Using rare loss-of-function variants, missense variants predicted to be pathogenic or likely pathogenic, and a subset of rare synonymous variants that cause large changes in local codon bias as independent data sets, we found strong, phenome-wide disease associations shared by two or more variant categories for 39% of the GPCRs. To validate the bioinformatics and sequence kernel association test analyses, we functionally characterized rare missense and synonymous variants of GPR39, a family A GPCR, revealing altered expression or Zn 2؉-mediated signaling for members of both variant classes. These results support the utility of rare variant analyses for identifying disease associations for GPCRs that lack impactful common variants. We highlight the importance of rare synonymous variants in human physiology and argue for their routine inclusion in any comprehensive analysis of genomic variants as potential causes of disease. The superfamily of G protein-coupled receptors (GPCRs) 3 translates extracellular signals from light, metabolites, and hormones into cellular changes, which makes them the targets of a significant fraction of drugs currently on the market (1). Genome-wide association studies (GWASs) on common variants in GPCRs have begun to identify their contributions to various disease processes (2, 3). However, many GPCRs lack functionally important common variants, and alternate strategies are needed to understand their roles in human physiology. Recently, sequence kernel association tests (SKATs) have been applied to rare variants to assess disease associations. Initial approaches binned all rare variants in a genomic region or gene (4, 5), but current methods group the rare variants most likely to contribute to disease associations or aggregate variants based on domain or family structures (6, 7). Large-scale studies of orphan or understudied GPCRs to characterize natural genetic variation in the human population can provide insights into the biological function and/or potential causal contributions to disease processes (8). The Dis-covEHR collaboration represents a tremendous resource (9) because whole exome sequences can be linked to the electronic health record (EHR) in a deidentified manner. Here we used whole exome sequences and clinical information from 51...