PROJECT SUMMARY/ABSTRACT Gestational diabetes mellitus (GDM), is among the most common pregnancy complications in the US. Shared pathophysiology with type-2 diabetes (T2D), evidence of familial aggregation, and evidence of racial disparity all support a role for genetic predisposition. Asian American (AAM) and Hispanic American (HA) women have lower prevalence of obesity on average than African American (AA) women, yet have higher GDM prevalence: 10.2%, 6.8%, 4.5% and 4.4% in AAM, HA, European American (EA) and AA women, respectively. Current studies are limited to candidate gene investigations with most investigating five to ten known T2D loci and only one genome-wide association study (GWAS) (468 cases; 1242 controls), in a South-Korean population. AAMs and HAs are the fastest growing populations in the US. Despite evident racial disparity suggesting a genetic etiology, no study has evaluated whether this is in part is rooted in differences associated with genetic ancestry. The overall goals of this proposal are to expand comprehensive genetic investigations of GDM and related traits by leveraging electronic health records (EHR) and bio-repositories to better understand the etiology which may inform personalized strategies for screening and prevention. We aim to develop, refine and validate reproducible and portable bioinformatics-algorithms to identity GDM cases and controls using de- identified EHR data at Vanderbilt. We will evaluate whether reported race/ethnicity modifies the association between maternal BMI and GDM in the Vanderbilt EHR database, the synthetic derivative (SD) (>8000 cases; Aim 1.1). In approximately 2,200 cases and 4,400 controls with genetic data, we will perform a Mendelian randomization study to test whether genetic instruments of central obesity (waist to hip ratio) or overall obesity (BMI) are more strongly associated with GDM (Aim 1.2). We will perform the first two-stage trans-ethnic GWAS of GDM in the US in EA, AA, HA, and AAM women from the SD and replicate associated variants (P < 1x10-6) in over 1000 GDM cases and many controls from the UK Biobank and Mount Sinai BioME EHR-linked bio- repository (Aim 2.1). By integrating GWAS data and expression quantitative trait loci (eQTL) data from various tissues with methods such as S-PrediXcan, we will prioritize candidate causal genes for GDM (Aim 2.2). Finally, we will explore whether genetically inferred Asian/Native American ancestry proportion is associated with increased risk of GDM (Aim 3.1) and T2D (Aim 3.2) in HAs. The well-tailored mentored training program supports the stated research aims and provides the candidate with the protected time to gain appropriate training in areas in which he lacks fully independent expertise, including phenotyping in the EHR setting, biomedical informatics and knowledge of gestational diabetes and classification of pregnancy outcomes. Successful completion of this award will facilitate the candidate's development into an independ...