UCLA

The biobank era of genomics has ushered in a multitude of opportunities for precision medicine research. In particular, biobanks connected to electronic health records (EHR) provide rich phenotype information used to study to clinical phenome. First, I describe two computational methods designed to infer the genetic architecture of complex traits using biobank-scale data. Both methods are based on Markov Chain Monte Carlo techniques. Next, I provide an overview of the UCLA ATLAS Community Health Initiative (ATLAS), an EHR-linked biobank embedded within UCLA Health. Using this data set, I explore the role of genetic ancestry in common disease risk across the UCLA patient population. Next, I include a review of how race, ethnicity, and genetic ancestry are utilized in the field of EHR- linked biobanks. Finally, I propose an EHR-based algorithm, called PheNet, which identifies undiagnosed patients with Common Variable Immunodeficiency Disorders and demonstrate its application across a total of 5 University of California Health systems.