Meta-Analysis of Transcriptome-Wide Association Studies across 13 Brain Tissues Identified Novel Clusters of Genes Associated with Nicotine Addiction.

Zhenyao Ye,Braxton D Mitchell,Chen Mo,Hongjie Ke,Tianzhou Ma,Shuo Chen,Peter Kochunov,Chixiang Chen,L Elliot Hong,Qi Yan

doi:10.3390/genes13010037

Abstract

Genome-wide association studies (GWAS) have identified and reproduced thousands of diseases associated loci, but many of them are not directly interpretable due to the strong linkage disequilibrium among variants. Transcriptome-wide association studies (TWAS) incorporated expression quantitative trait loci (eQTL) cohorts as a reference panel to detect associations with the phenotype at the gene level and have been gaining popularity in recent years. For nicotine addiction, several important susceptible genetic variants were identified by GWAS, but TWAS that detected genes associated with nicotine addiction and unveiled the underlying molecular mechanism were still lacking. In this study, we used eQTL data from the Genotype-Tissue Expression (GTEx) consortium as a reference panel to conduct tissue-specific TWAS on cigarettes per day (CPD) over thirteen brain tissues in two large cohorts: UK Biobank (UKBB; number of participants (N) = 142,202) and the GWAS & Sequencing Consortium of Alcohol and Nicotine use (GSCAN; N = 143,210), then meta-analyzing the results across tissues while considering the heterogeneity across tissues. We identified three major clusters of genes with different meta-patterns across tissues consistent in both cohorts, including homogenous genes associated with CPD in all brain tissues; partially homogeneous genes associated with CPD in cortex, cerebellum, and hippocampus tissues; and, lastly, the tissue-specific genes associated with CPD in only a few specific brain tissues. Downstream enrichment analyses on each gene cluster identified unique biological pathways associated with CPD and provided important biological insights into the regulatory mechanism of nicotine dependence in the brain.

Highlights

The past decade has witnessed an explosion in genome-wide association studies (GWAS) research, which identified thousands of robust, reproducible genetic risk variants associated with complex diseases and traits [1,2]
The meta-analysis of TS-transcriptome-wide association studies (TWAS) identified 48 genes significantly associated with cigarettes per day (CPD) at false discovery rate (FDR) < 0.05 common in both UK Biobank (UKBB) and GSCAN
Comparing to S-MultiXcan and tissue-specific TWAS (TS-TWAS), meta-analysis was overall more powerful in identifying more nicotine-addiction-associated genes (Table 1; Figure 2 highlighted in red, Figure S1), especially among genes with heterogeneous association patterns across tissues (Figure S1, Supplementary file 1)

Summary

Introduction

The past decade has witnessed an explosion in genome-wide association studies (GWAS) research, which identified thousands of robust, reproducible genetic risk variants associated with complex diseases and traits [1,2]. The loci identified by GWAS are not directly interpretable due to the strong linkage disequilibrium (LD) that obscures the causal variants, and GWAS data alone can hardly determine the causal genes and the underlying regulatory mechanism [5]. To fill this gap, transcriptome-wide association studies (TWAS) are developed to utilize expression quantitative trait loci (eQTL) cohorts (e.g., Genotype-Tissue Expression (GTEx) [6]), which include both genotype and gene expression data as a reference panel to infer association with a trait at the gene level [7].

Objectives

Methods

Results

Conclusion