GENOME WIDE ASSOCIATION STUDY META-ANALYSIS OF HOMOARGININE USING THE HRC REFERENCE PANEL

Abstract

Objective: Data from the LUdwigshafen RIsk and Cardiovascular Health (LURIC) study and the 4D Study have shown that low homoarginine (HARG) is a significant risk factor for all-cause and cardiovascular mortality. In 2013, we performed a first genome wide association study (GWAS) meta-analysis on the genetic determinants of homoarginine concentration based on data from LURIC and the Young Finns study (YFS) and discovered three loci with genome-wide significance. In the meantime, new expanded and more accurate imputation reference panels have been released. Therefore, our aim was to performed a meta-analysis of GWAS using the haplotype reference consortium (HRC) imputation reference panel to identify further genes which are involved in the regulation of HARG serum levels. Furthermore, we increased the sample size from ca. 5k to 6k. Design and method: HARG was measured in serum stored at -80°C by a reversed phase high-performance liquid chromatography. Genotyping was done in LURIC and 4D by using the Affymetrix 6.0 array and a custom-built Illumina Human 670k BeadChip in YFS. Imputation to the HRC reference panel was performed using Minimac. SNPs were evaluated for association with HARG using linear regression analyses using the software PLINK2 with adjustment for age, sex and principle components. Meta-analysis was done using a fixed-effects, effective sample-weighted Z-score meta-analysis method, as implemented in the software METAL. Results: In the meta-analysis, 391 SNPs were genome-wide significantly associated with HARG. These SNPs are located on chromosome 2 (CPS1), on chromosome 5 (AGXT2), on chromosome 6 (ARG1) and on chromosome 15 (GATM) (Fig. 1). This adds one significant locus to the previous meta-analysis. Another six loci showed suggestive association with P-values < 10E-6. Gene set enrichment and tissue enrichment analyses using DEPICT did not results in any significant hits after FDR correction. Highest-ranking gene sets with p-values of 10E-4 were gene sets involved in dystonia, cellular hormone response, insulin resistance, carbohydrate and lipid metabolism. Highest-ranking tissues with nominally significant p-values were atrial appendage and heart atria. Conclusions: In the meta-analysis four genomic loci reached genome-wide significance and a further six loci showed suggestive association.