Integrating expression-related SNPs into genome-wide gene- and pathway-based analyses identified novel lung cancer susceptibility genes.

Traditional pathway analysis map single nucleotide polymorphisms (SNPs) to genes according to physical position, which lacks sufficient biological bases. Here, we incorporated genetics of gene expression into gene- and pathway-based analysis to identify genes and pathways associated with lung cancer risk. We identified expression-related SNPs (eSNPs) in lung tissues and integrated these eSNPs into three lung cancer genome-wide association studies (GWASs), including 12,843 lung cancer cases and 12,639 controls. We used SKAT-C for gene-based analysis, and conditional analysis to identify independent eSNPs of each gene. ARTP algorithm was used for pathway analysis. A total of 374,382 eSNPs in the GWAS datasets survived quality control, which were mapped to 5,084 genes and 2,752 pathways. In the gene-based analysis, nine genes showed significant associations with lung cancer risk. Among them, TP63 (3q28), RP11-650L12.2 (15q25.1) and CHRNA5 (15q25.1) were located in known lung cancer susceptibility loci. We also validated two newly identified susceptibility loci (RNASET2 and AL133458.1 in 6q27, and MPZL3 in 11q23.3). Besides, DVL3 (3q27.1), RP11-522I20.3 (9q21.32) and CCDC116 (22q11.21) were identified as novel lung cancer susceptibility genes. Pathway analysis showed that pathways involved in protein structure, the Notch signaling pathway and the nicotinic acetylcholine receptor-related pathways were associated with lung cancer risk. Combing eSNPs, gene- and pathway-based analysis identifies novel lung cancer susceptibility genes, which serves as a powerful approach to decipher biological mechanisms underlying GWAS findings.