Identifying expression and DNA methylation biomarkers for lung adenocarcinoma risk in East Asia [RNA-seq]

Lung adenocarcinoma (LUAD) is the most common type of lung cancer. A recent genome-wide association study (GWAS) of LUAD in East Asia reported 28 independent susceptibility variants across 25 loci and identified 2 genes whose genetically predicted expression levels are associated with LUAD risk, using an ancestry-matched lung tissue expression quantitative trait loci (eQTL) dataset. It is desirable to identify additional susceptibility loci and to understand their underlying biological mechanisms. Using an expanded GWAS of LUAD in East Asia and ancestry-matched lung tissue eQTL and DNA methylation QTL datasets, we performed transcriptome-wide association studies and DNA methylome-wide association studies simultaneously and examined the association between measured expression of genes and DNA methylation of nearby CpGs (eQTM). Genes and nearby CpGs are termed CpG–gene–LUAD trios if these three associations hold simultaneously. At Bonferroni-corrected P<0.05, we identified a new susceptibility locus (6p21.31; lead SNP rs7772643), 10 LUAD-associated genes and 86 LUAD-associated CpGs. At false discovery rate q<0.05, we identified 28 LUAD-associated genes, 220 LUAD-associated CpGs, and 45 CpG–gene–LUAD trios; among them, 43 were direction-matched regarding these three associations. These show that 23 of the known 28 susceptibility variants for LUAD in East Asia are near these genes or CpGs and MARCH3, ELF5, IKZF3, GSDMB, CCDC116, and DSP are putative novel susceptibility loci. Few of them was reported in LUAD in European populations. This study substantially advances our understanding of the etiology of LUAD in East Asia and could be useful in developing translational applications.