Abstract 13286: Finding the Missing Heritability of Serum Lipids: Leveraging the Relatedness in a Large-scale Biobank

Abstract

Introduction: High prevalence of dyslipidemia in the United States results in elevated risk of cardiovascular disease. Although the high heritability of serum lipids has motivated many genome-wide association studies (GWAS), much of the heritability is still unexplained. Genomic regions shared with identity by descent (IBD) are inherited from the same common ancestor without recombination, and can be observed in both close and distant relatives. IBD segments provide an opportunity to discover genes that harbor low frequency, large effect variants that are undetectable in GWAS due to low power at rare and heterogeneous alleles. Methods: The Vanderbilt biobank (BioVU) comprises over 280,000 DNA samples from participants with linked electronic medical record (EMR). In this study, we extracted 18,337 European dyslipidemia cases, who have either diagnosed dyslipidemia or record of taking lipid-lowering drugs in linked EMR, and 18,337 sex, age, and ancestry-matched controls. The pairwise IBD shared segments were identified by GERMLINE using genotype data. To assess enrichment of IBD, we compared local shared IBD rates in case-case pairs and case-control pairs. Results: The greatest enrichment of IBD sharing was found on chromosome 5 (chr5: 169-171 MB, p-value=3.3х10 -5 ), and another three suggested regions were from chromosome 10, 18 and 22 (chr10: 86-87 MB, p-value=5.5х10 -4 , chr18: 10-11 MB, p-value=5.3х10 -4 , and chr22: 49.3-49.7 MB, p-value=2.1х10 -4 ). To further explore the molecular function, we used PrediXcan to impute the genetically regulated expression of the genes which locate within 1 MB of identified regions. A total of 31 unique genes in these regions exhibit significantly different predicted expression between risk IBD segment carriers and non-carriers, and are also significantly correlated with at least one serum lipids level. For instance, a novel long non-coding RNA, LINC01310, from chr22 associated with total glyceride (p-value=5.3х10 - 3 ) and expressed a lower level in risk people carrying the risk-associated IBD segment (p-value=1.2х10 - 6 ). Conclusions: Our results identity four genomic segments enriched in dyslipidemia cases, and, using GReX, we demonstrate the potential function of novel genes on the regulation of serum lipids.