Integrated characterisation of cancer genes identifies key molecular biomarkers in stomach adenocarcinoma

Aims Gastric cancer is one of the leading causes for cancer mortality. Recent studies have defined the landscape of genomic alterations of gastric cancer and their association with clinical outcomes. However, the pathogenesis of gastric cancer has not been completely characterised. Methods Driver genes were detected by five computational tools, MutSigCV, OncodriveCLUST, OncodriveFM, dendrix and edriver, using mutation data of stomach adenocarcinoma (STAD) from the cancer genome altas database, followed by an integrative investigation. Results TTN, TP53, LRP1B, CSMD3, OBSCN, ARID1A, FAT4, FLG, PCLO and CSMD1 were the 10 most frequently mutated genes. PIK3CD, NLRC3, FMNL1, TRAF3IP3 and CR1 were the top five hub genes of the blue coexpression module positively correlated with pathological tumour stage and lymph node stage (p values <0.05 for all cases). Hierarchical clustering analysis of copy number variations of driver genes revealed three subgroups of STAD patients, and cluster 2 tumours were significantly associated with lower lymph node stage, less number of positive lymph nodes and higher microsatellite instability and better overall survival than cluster 1 and cluster 3 tumours (p values <0.05 for all cases, Wilcoxon rank-sum test or log rank test). High expression in one or more of DNER, LHCGR, NLRP14, OR4N2, PSG6, TTC29 and ZNF568 genes was associated with increased mortality (p values <0.05 for all cases, log rank test). Conclusions The driver genes shed insights into the tumourigenesis of gastric cancer and the genes DNER, LHCGR, NLRP14, OR4N2, PSG6, TTC29 and ZNF568 pave the way for developing prognostic biomarkers for the disease.