When Genomes Join the Data Ecosystem: New Horizons from All of Us Multimodal Research

Presentation Type: Panel
Presentation Time: 11:15 AM - 12:30 PM

Primary Track: Translation Bioinformatics/Precision Medicine


The All of Us Research Program provides an unprecedented opportunity to accelerate precision medicine across populations by integrating comprehensive phenotype data including electronic health records (EHR) with large-scale whole-genome sequencing. Realizing the scientific value of this multimodal dataset requires advanced informatics methods capable of harmonizing genomic and clinical information and supporting scalable analysis. This panel brings together experts developing complementary capabilities in structural variant imputation, computational phenotyping, multimodal large language foundation models integrating genomic data, and cloud-native approaches. Presentations will highlight innovations in structural variant imputation using long-read sequencing data, phenotype risk scores for genomic discovery, transformer-based EHR-genomic modeling, and high-performance computational infrastructures supporting genomic and phenomic analyses in All of Us. Collectively, these talks outline a cohesive methodological framework for conducting rigorous and efficient genomic research. Attendees will gain insight into the analytic strategies necessary to transform large-scale multimodal data into clinically meaningful knowledge.

Moderator:
Chenjie Zeng, PhD
NIH

Speaker(s):
Henry Taylor, PhD
NIH

Alessandra Breschi, PhD
Verily

Lisa Bastarache
Vanderbilt University Medical Center

Ariel Williams, PhD
University of Virginia

Author(s):
Henry Taylor, PhD - NIH; Lisa Bastarache - Vanderbilt University Medical Center; Alessandra Breschi, PhD - Verily; Tam Tran; Chenjie Zeng, PhD - NIH;

---