Abstract 12478: Investigating the Shared Genetic Architecture Between Coronary Artery Disease and Major Depressive Disorder

Abstract

Coronary Artery Disease is a highly polygenic disorder and has increased comorbidity with Major Depressive Disorder (MDD). Our previous work demonstrated that the PRS for MDD was significantly associated with manifestations of CAD including ischemic heart disease (OR = 1.09, p=8.28x10-6) and coronary atherosclerosis (OR = 1.10, p=1.08x10-5) in a large hospital population. Furthermore, the association remained robust after adjusting for known CAD risk factors, suggesting that some proportion of genes that increase susceptibility to MDD act pleiotropically to increase CAD risk. Therefore, the aim of our study was to identify genes that increase risk for both CAD and MDD. We first performed an imputed TWAS using existing GWAS data on CAD from the CARDIoGRAMplusC4D consortium and UKBB and MDD from Psychiatric Genomics Consortium. TWAS uses the summary statistics from a GWAS and expression prediction models to infer the association between the genetically predicted expression of each gene and a phenotype of interest. We identified genes for which predicted expression was associated with increased risk of both MDD and CAD. We then further tested the relationship between those genes and quantitative biomarkers recorded in the Vanderbilt electronic health record (EHR). For this, we used Vanderbilt’s collection of de-identified biosamples (BioVU) which currently includes ~100k genotyped patient DNA samples linked with their EHRs. We used PredixVU, a gene-based association method to determine whether genes identified in the TWAS were associated with clinical biomarkers. We identified significant overlap in the genes associated with both MDD and CAD at an FDR of 0.01 (hypergeometric enrichment factor of 3.2; p = 0.002) and 0.05 (hypergeometric enrichment factor of 2.4; p = 8.35e-9). TMEM106B was significantly associated with both MDD and CAD after Bonferroni correction was applied. TMEM106B is involved in lysosomal activity and has been independently associated with both, CAD and MDD. The predicted expression of TMEM106B in the putamen is associated with decreased levels of hydroxyproline (OR = 0.830, P-Value = 7.5 e-5). For our future work, we aim to investigate the shared pathways of these genes across Asian, European and African ancestry using BioVU.