Convergence of biomarkers and risk factor trait loci of coronary artery disease at 3p21.31 and HLA region

Majid Nikpay,Ruth Mcpherson

doi:10.1038/s41525-021-00174-z

Majid Nikpay, Ruth Mcpherson

Open Access

https://doi.org/10.1038/s41525-021-00174-z

Copy DOI

Abstract

Here we seek to identify molecular biomarkers that mediate the effect of risk factors on coronary artery disease (CAD). We perform a SNP-based multiomics data analysis to find biomarkers (probes) causally associated with the risk of CAD within known genomic loci for its risk factors. We identify 78 biomarkers, the majority (64%) of which are methylation probes. We detect the convergence of several CNS and lifestyle trait loci and their biomarkers at the 3p21.31 and human leukocyte antigen (HLA) regions. The 3p21.31 locus was the most populated region in the convergence of biomarkers and risk factors. In this region, we noted as the BSN gene becomes methylated the level of stomatin (STOM) in blood increases and this contributes to higher risk of CAD. In the HLA locus, we identify several methylation biomarkers associated with various CAD risk factors. SNPs in the CFB gene display a trans-regulatory impact on the GRIA4 protein level. A methylation site upstream of the APOE gene is associated with a higher protein level of S100A13 which in turn leads to higher LDL-C and greater CAD risk. We find UHRF1BP1 and ILRUN mediate the effect of obesity on CAD whereas methylation sites within NOS3 and CKM mediate the effect of their associated-risk factors on CAD. This study provides further insight into the biology of CAD and identifies a list of biomarkers that mediate the impact of risk factors on CAD. A SNP-based initiative can unite data from various fields of omics into a single network of knowledge.

Highlights

GWAS studies have identified numerous loci associated with various complex phenotypes including the susceptibility to coronary artery disease (CAD)
Using our SNP-based multiomics data analysis plan (Fig. 1), we identified 78 biomarkers associated with various risk factors for CAD at GWAS significance level (P < 5e−8) as well as to CAD per se after correction for multiple testing (Supplementary Data 3)
After relaxing our linkage disequilibrium (LD) threshold to r2 < 0.2, we found a higher level of SELE is causally contributing to lower risk of CAD [B = −0.05, P = 5e−13, Number of SNPs (NSNPs) = 6] using pQTLs from this study

Summary

Introduction

GWAS studies have identified numerous loci associated with various complex phenotypes including the susceptibility to coronary artery disease (CAD). We have GWAS data that catalog the associations between genome and the phenome but cannot provide molecular insight and we have GWAS data from omics studies that report associations between genome and various molecular features. By combining these two different sets of data, it is possible to identify genomic regions where SNPassociation signals are consistent (co-localize) for a trait and a molecular feature (biomarker), Mendelian randomization (MR) can be used to test whether change in the level of the biomarker is causally contributing to the trait (Fig. 1 and Supplementary Fig. 1). We have extended these studies to identify genomic loci through which CAD risk factors exert their effects

Objectives

Methods

Results

Conclusion