Evergreen Ballroom Prefunction
Imputed Transcriptome-Wide Association Studies of Blood Cell Traits in European Ancestry Cohorts (306698)
Eric Boerwinkle, University of Texas Health Science CenterDavid Couper, University of North Carolina
Santhi Ganesh, University of Michigan
Misa Graff, University of North Carolina
Megan Grove, University of Texas Health Science Center
Eric Jorgenson, Kaiser Permanente Northern California
Yun Li, University of North Carolina
Alanna Morrison, University of Texas Health Science Center
Kari North, University of North Carolina
Laura Raffield, University of North Carolina
Alexander P Reiner, University of Washington
Jonathan Rosen, University of North Carolina
*Amanda L Tapia, University of North Carolina
Kristin Young, University of North Carolina
Bing Yu, University of Texas Health Science Center
Keywords: TWAS, blood cell traits, penalized regression, imputation
Blood cell traits (BCTs) are important intermediate clinical phenotypes for many diseases and are highly heritable. Although genome-wide association studies (GWAS) have identified >2700 variants for BCTs, the biological mechanisms underlying these associations remain largely unknown. To better understand these mechanisms, we aim to associate imputed transcript levels with BCTs in individuals of European ancestry in the Atherosclerosis Risk in Communities Study, Genetic Epidemiology Research on Adult Health and Aging, and UK Biobank. We follow a transcriptome-WAS (TWAS) approach using penalized regression to first build prediction models in a training set (e.g. GTEx whole blood). With the trained models, we then impute transcripts in each individual and associate imputed transcripts with BCTs. The TWAS approach reduces multiple testing burden by assessing variants’ aggregate effect on transcripts, allows for better identification of the target gene at known loci for BCTs, and can identify novel loci missed by single variant analyses. These ongoing TWAS analyses involve >550,000 individuals and are well-powered to explore the transcriptional impact of BCT-associated genetic variants.