Abstract 11596: Rare Coding Variation and Risk for Myocardial Infarction: an Exome Chip Study of ∼6,000 Cases and Controls
Myocardial infarction (MI) is the leading cause of death in the U.S. and is heritable. Through genome-wide association studies (GWAS), the role of common genetic variation in risk of MI has been thoroughly evaluated and more than 30 loci have been associated with MI. The extent to which rare variants contribute to MI risk is unknown. We tested the hypothesis that rare coding variation contributes to MI risk. We evaluated 1,989 cases with MI and 3,995 controls free of MI from two cohorts [[Unable to Display Character: &–]] the Ottawa Heart Study and the Women's Health Initiative [[Unable to Display Character: &–]] and genotyped these samples using a novel genotyping array (“exome chip”) focused on genetic variation in the protein coding regions of the human genome. The “exome chip” contains ∼250,000 coding variants discovered through exome sequencing in ∼12,000 individuals and, in addition, includes all common GWAS variants previously associated with MI. Collectively, the array represents nearly all non-synonymous coding and splice-site variation with a >1:1000 allele frequency in the European population. Association testing for MI was performed using single-marker logistic regression correcting for ten principal components of ancestry. From this screen, the strongest association signal was seen at the common non-coding variant at chromosome 9p21 previously discovered by GWAS (55% allele frequency, OR 1.27, p=3x10-9). For rare coding variants (defined as an allele frequency < 5%), the strongest association signal was with LPA (I4399M, 2.3% allele frequency, OR 1.73, p=1x10-5). Of note, a well-studied variant in PCSK9 (R46L, 1.6% frequency, OR = 0.66) had a modest association (p=0.02). After excluding previously reported GWAS variants, there is an excess of rare coding variants with p<0.05 but which do not exceed chip-wide significance (p<2x10-7). Post-hoc power calculations show that large sample sizes will be required to confidently detect the signal from rare variants. For example, 13,500 cases and 13,500 controls would be required to detect a locus with the same frequency and effect as PCSK9 R46L at chip-wide significance. At present, we are attempting to replicate our top novel results in additional independent samples to provide more refined insights into the role of rare coding variation in risk for MI.
- © 2012 by American Heart Association, Inc.