Coronary artery disease (CAD) is a leading cause of death and disability worldwide and represents a common complex disease with genetic and environmental determinants. Genome-wide association studies (GWAS) of CAD from international consortia, including the CARDIoGRAMplusC4D Consortia, and analysis of UK Biobank data have identified over 150 independent variants associated with the risk of CAD. In this study, we leverage multiple sources of molecular 'omics data from the NHLBI Trans-Omics for Precision Medicine (TOPMed) program to examine prior findings with (a) overlap of GWAS with eQTL/pQTL, and (b) network structure. Our study leveraged transcriptomic data obtained from peripheral blood mononuclear cells (PBMCs) and plasma proteomics from participants in the Multi-Ethnic Study of Atherosclerosis (MESA). We performed Bayesian colocalization using R/coloc to identify the molecular targets implicated by GWAS. Support for hypothesis 4 (both the QTL and GWAS are associated with the region and share a single causal variant) used a posterior probability threshold of 0.8 for Bayesian colocalization. To investigate the relationship of the colocalized genes and proteins with subclinical atherosclerosis, we conducted Bayesian colocalization using the results of GWAS for coronary artery calcium (CAC) and carotid artery intima-thickness (cIMT), focusing on those genes identified in our initial analyses of CAD. We performed weighted gene co-expression network analyses (WGCNA) to identify the modules of highly correlated genes and calculated the module membership (MM threshold of 0.8) to identify the hub genes for each module. We examined the association of the hub genes with CAC in MESA using linear regression with adjustment for age, sex, and study site to identify the CAC-related hub genes for each module. Based on our Bayesian colocalization, we identified 224 putative genes and 78 putative proteins by integrating eQTL and pQTL from TOPMed MESA with summary statistics from the prior GWAS of CAD (van der Harst and Verwejj, 2018). Among these, we observed additional colocalization with subclinical atherosclerosis traits for ILRUN (cIMT), GGCX (CAC), and RNF114 (CAC). Using WGCNA on the TOPMed RNA-seq data from MESA PBMCs, we constructed 30 modules of highly correlated genes and further identified the CAC-related hub genes for each module. We found AHCYL1 to be a CAC-related hub gene from the “dark green” module, also located within the region of the lead variant on chromosome 1 from the GWAS of CAD. Our study identified multiple candidate genes at GWAS loci that are supported by molecular QTL and further showed the value of network analysis to highlight the core role of some genes implicated by GWAS.

Coronary artery calcification (CAC) is a marker of atherosclerosis and is associated with increased risk of coronary heart disease (CHD) mortality, especially in individuals with type 2 diabetes (T2D). While numerous studies have identified genetic loci involved in CAC, CHD, and T2D, the shared genetic architecture between these highly associated traits is still being understood. We compared the effects of 388 genetic variants significantly associated with CHD in the literature in 2,971 individuals with T2D and 13,022 normoglycemic controls utilizing whole genome sequencing generated by the National Heart, Lung, and Blood Institute’s Trans-Omics for Precision Medicine (TOPMed) program. Participants were from four race/ethnic groups, including European American, African American, Hispanic/Latinx, and East Asian. CAC was first log transformed, then further transformed through inverse rank-based normalization of the residuals accounting for age and sex. Linear mixed models accounting for relatedness, implemented in GENESIS, were used to test for interaction between each variant and T2D status. Analyses were adjusted for age, sex, study and the first eleven principal components. The genetic main and interaction effects were assessed in a joint test using a two degree of freedom model to determine if a CHD variant was associated with CAC, then further evaluated to determine if these variants had a significantly different effect in T2D cases versus controls. Using Bonferroni corrected significance threshold of P<1.3x10-4 (0.05/388), we identified 20 CHD variants associated with CAC according to the joint test, of which 11 had a statistically significant different effect in T2D cases and controls (rs1807214 near ABHD2, rs668948 near APOB, rs11655024 in BCAS3, rs840616 near CALCRL, rs1321309 near CDNK1A, rs6883598 near FBN2, rs249760 near FGF1, rs12897 near FNDC3B, rs12691049 near MYH11, rs6494488 near RBPMS2, rs7118294 near WT1). While rs668948 has not previously been implicated in previous CAC GWAS, it lies nearest to APOB, which is a known driver of plaque development and subsequent atherosclerosis. Similarly, rs840616 lies near CALCRL, which is involved in the maintenance of calcium homeostasis. Overall, 159 of the CHD variants were nominally significant (P<0.05) for CAC according to the joint test, including 96 variants with nominally significant T2D interactions. These results highlight T2D as an important moderator of the association of CHD loci with subclinical atherosclerosis.

Coronary artery calcification (CAC) is a marker of atherosclerosis and is associated with increased risk of coronary heart disease (CHD) mortality, especially in individuals with type 2 diabetes (T2D). While numerous studies have identified genetic loci involved in CAC, CHD, and T2D, the shared genetic architecture between these highly associated traits is still being understood. We compared the effects of 388 genetic variants significantly associated with CHD in the literature in 2,971 individuals with T2D and 13,022 normoglycemic controls utilizing whole genome sequencing generated by the National Heart, Lung, and Blood Institute’s Trans-Omics for Precision Medicine (TOPMed) program. Participants were from four race/ethnic groups, including European American, African American, Hispanic/Latinx, and East Asian. CAC was first log transformed, then further transformed through inverse rank-based normalization of the residuals accounting for age and sex. Linear mixed models accounting for relatedness, implemented in GENESIS, were used to test for interaction between each variant and T2D status. Analyses were adjusted for age, sex, study and the first eleven principal components. The genetic main and interaction effects were assessed in a joint test using a two degree of freedom model to determine if a CHD variant was associated with CAC, then further evaluated to determine if these variants had a significantly different effect in T2D cases versus controls. Using Bonferroni corrected significance threshold of P<1.3x10-4 (0.05/388), we identified 20 CHD variants associated with CAC according to the joint test, of which 11 had a statistically significant different effect in T2D cases and controls (rs1807214 near ABHD2, rs668948 near APOB, rs11655024 in BCAS3, rs840616 near CALCRL, rs1321309 near CDNK1A, rs6883598 near FBN2, rs249760 near FGF1, rs12897 near FNDC3B, rs12691049 near MYH11, rs6494488 near RBPMS2, rs7118294 near WT1). While rs668948 has not previously been implicated in previous CAC GWAS, it lies nearest to APOB, which is a known driver of plaque development and subsequent atherosclerosis. Similarly, rs840616 lies near CALCRL, which is involved in the maintenance of calcium homeostasis. Overall, 159 of the CHD variants were nominally significant (P<0.05) for CAC according to the joint test, including 96 variants with nominally significant T2D interactions. These results highlight T2D as an important moderator of the association of CHD loci with subclinical atherosclerosis.