Integrative analysis of eQTL and GWAS summary statistics reveals novel genes related to Alzheimer's disease

Abstract

Listen

Translate article

AbstractBackgroundLarge‐scale genome wide association studies (GWASs) have identified many genetic variants associated with Alzheimer’s disease (AD) and related traits. While many GWAS findings were found to co‐localize with expression quantitative trait loci (eQTL), it is of great interest to investigate the functional effect of GWAS hits on the downstream transcriptomic level. Considering the lack of gene expression data from the brain tissue, we propose to integrate the summary statistics from AD GWAS and brain eQTL analysis to investigate potential transcriptomic alterations inside AD brains. We further validate our findings using the brain gene expression data from the ROS/MAP cohort.MethodWe used the GWAS summary statistics from IGAP[1] and brain eQTL results from BRAINEAC[2]. Summary Mendelian Randomization (SMR) and Heterogeneity in Dependent Instruments (HEIDI) tests were applied to identify highly significant genes related to AD in temporal cortex, frontal cortex and hippocampal regions. For significant genes identified from each region, we further performed differential gene expression analysis using the RNA‐Seq data from the corresponding brain tissue in the Mayo Clinic cohort. SNPs from these genes were tested for association with FDG intensity and thickness of corresponding brain regions using the data from ADNI. Finally, we performed pathway analysis using ClueGO.ResultSMR and HEIDI test identified 32 genes significantly associated with AD in the transcriptomic level, but only in temporal cortex. 10 of them were further validated with altered expression in AD temporal cortex region in the Mayo cohort (Fig.1). 19 SNPs from these genes were significantly related to the FDG intensity and thickness of temporal cortex regions, predominantly present in TOMM40 and NECTIN genes. Top pathways enriched by these genes are Neutrophil degranulation and cell surface interactions at vascular wall (Fig.2).ConclusionWe identified several SNPs with potential regulatory role in mediating the expression level of genes, which are found altered in AD brains and associated with multiple neuroimaging phenotypes. With evidence from multiple sources, these SNPs, their downstream genes and related pathways could serve as potential targets for further therapeutic intervention of AD.