Abstract 16492: A Study of 63,253 Cases and 126,820 Controls Identifies Multiple Novel Loci for Coronary Artery Disease and Detects Independent Signals in Many of the Established Ones
Genome-wide association studies have identified 30 loci conferring risk to coronary artery disease (CAD) and myocardial infarction (MI) but in most cases the causative variants remain unknown. We set an experiment to fine map 22 of these established loci and in parallel to evaluate a large number of untested associations identified in GWA meta-analyses. The Metabochip array is a custom iSELECT chip containing 196,725 SNPs, designed for purposes (1) follow-up of putative associations in several cardiometabolic traits and (2) fine mapping of confirmed loci in these traits. In the CARDIoGRAMplusC4D consortium we undertook a two stage meta-analysis to test for disease risk 5658 SNPS associated with CAD at P<0.01 in CARDIoGRAM (22,233 cases and 64,762 controls) discovery phase. Stage 2 involved 34 studies totaling 41,020 cases and 62,058 controls of European or South Asian descent (including GWA studies imputed with HapMap). We applied study-wise standard QC and corrected for population stratification. We then combined p-values for all 5658 SNPs from Stage 1 with their respective Stage 2 p-values using Fisher's method. We discovered 12 novel loci reaching genome wide significance (WGS; p<5 x 10-8): IL6R, APOB, VAMP8-GGCX, SLC22A4, GUCY1A3, KCNK5, PLG, TRIB1, ABCG5, FURIN, and FLT1. Four of the novel WGS CAD loci APOB, TRIB1, ABCG5, and FURIN are known lipid loci. Initial fine mapping analysis in the 22 CAD loci was carried out only in samples typed with the Metabochip which includes low frequency variants (14% of markers have MAF < 0.02 ) - integration of GWA studies is underway using imputation to the 1000 Genomes project. Single marker analysis has identified a new lead SNP in 19 of the 22 loci. Many signals localized to much tighter intervals than the corresponding recombination blocks. Interestingly, in five loci we also detected an independent signal to the previously reported one (index SNP). For example, in the COL4A1/A2 locus the new independent signal is stronger (OR=0.91, P=1.19x10-08) compared to the known one (OR=0.94, P=2.23x10-04). The 13 new loci reported here increase our insight in to the pathogenesis of CAD and initial fine mapping results suggest a complex genetic architecture of risk regions paving the way towards elucidating causal disease mechanisms.
- © 2011 by American Heart Association, Inc.