Abstract 12914: Explaining the Heritability in Lipoprotein Subclasses Using Detailed Genetic Information in Known Lipid Loci
Genome-wide association studies (GWAS) have identified over 100 loci that are associated with TC, TG, LDL-C, HDL-C or lipoprotein subclasses, and further fine-mapping studies have shown that many of the loci carry multiple independently associated variants. We set to test the total contribution of the associated SNPs on the variation of lipoprotein subclasses. 14 lipoprotein subclasses with their lipid and particle concentrations were measured using NMR in 4703 individuals from the Northern Finland Birth Cohort 1966. The proportion of variance explained was studied using SNP scores built from the known GWAS SNPs (N=108) alone and with the known additional independently associated SNPs in the same loci (N=16). Together the reported GWAS SNPs explained 4.5-10.5% of variance in the lipoprotein subclasses; IDL, very-small VLDL and large HDL measures showed the highest and small HDL measures the lowest explained variances. The difference between the HDL particles suggests that the currently known genetic determinants for HDL mainly reflect the variance arising from the larger HDL particles. As many of the lipoprotein loci have been shown to carry multiple independently associated SNPs, we estimated how much these variants jointly with the GWAS SNPs explain of the variance in the subclass measures. The proportion increased substantially, especially for the very-large HDL measures. For example, for the triglyceride content of the very-large HDLs the explained proportion increased from 10.3% to 16.0%, or 25% of the trait heritability. Interestingly, much of this increase is attributable to a single locus, hepatic lipase (LIPC), which harbors four independent previously associated variants. In conclusion, as yet the known SNPs unevenly capture the full spectrum of lipoprotein variance, mostly focusing on very-small VLDL, IDL and large HDL. This heterogeneity may in part explain why the lipid SNPs modify the CHD risk very differently. Also, the results highlight that with current lipid genes we explain only a modest fraction of variability in some lipoprotein measures, especially for small HDL. Discovery of further loci and better coverage of genetic variability in known loci is likely to explain a bigger portion of the differences in lipoprotein profiles.
- © 2012 by American Heart Association, Inc.