Population-scale Long-read Sequencing in the All of Us Research Program

Area of Science:

• Genomics and Bioinformatics
• Human Genetics
• Precision Medicine

Background:

• The All of Us Research Program (AoU) aims to build a national biobank linking genomic data with electronic health records (EHRs).
• Short-read whole-genome sequencing (srWGS) has limitations in detecting complex structural variations (SVs).

Purpose of the Study:

• To present the first large-scale analyses of long-read sequencing (LRS) in the AoU.
• To develop a framework for deriving genomic insights into SVs relevant to human health and disease.
• To impute SVs into existing srWGS datasets for trait association studies.

Main Methods:

• Joint analysis of LRS data from 1,027 self-identified Black or African American individuals using Pacific Biosciences HiFi technology.
• Development of cohort-level variant calling and a scalable workflow for SV imputation.
• Analysis of SV-disease associations in 10,000 AoU participants with srWGS and matched EHRs.

Main Results:

• A comprehensive variant callset including repeat expansions, clinically relevant haplotypes (e.g., HLA, CYP2D6), and SVs was constructed from LRS data.
• 291 SV-disease associations across 226 conditions were identified, with 50.9% involving SVs absent in srWGS.
• Fine-mapping identified 191 SV-disease pairs where SVs were the strongest drivers of association, particularly in individuals with African ancestry.

Conclusions:

• LRS provides critical genomic insights into complex SVs missed by srWGS, impacting the understanding of human health and disease.
• The developed imputation framework enables scalable SV analysis in large cohorts.
• Integration of LRS into biobanks like AoU is transformative for precision medicine, especially for ancestries underrepresented in genomic studies.