Ruowang Li, PhD: “Genomic Discovery and Predictive Modeling with Biobank-Linked EHR Data”

The growing availability of biobank-linked electronic health records (EHR) provides unprecedented opportunities for genomic discovery and risk prediction. In this talk, I will highlight several recent methodological advances from our group that leverage large-scale EHR and genomic data. These include federated learning approaches for distributed analysis across health systems, transfer-learning methods for improving polygenic risk prediction, and a general framework for variable selection in high-dimensional data. Using examples from multiple biobank-linked EHR cohorts, I will illustrate how these methods enable more powerful, scalable, and generalizable genomic analyses.