LAUSR

Gastric cancer (GC) is one of the most fatal cancers in the world. Thousands of biomarkers have been explored that might be related to survival and prognosis via database mining. However, the prediction effect of single gene biomarkers is not specific enough. Increasing evidence suggests that gene signatures are emerging as a possible better alternative. We aimed to develop a novel gene signature to improve the prognosis prediction of GC. Using the messenger RNA (mRNA)‐mining approach, we performed mRNA expression profiling in a large GC cohort (n = 375) from The Cancer Genome Atlas (TCGA) database. Gene Set Enrichment Analysis (GSEA) was performed, and we recovered genes related to the G2/M checkpoint, which we identified with a Cox proportional regression model. We identified a set of five genes (MARCKS, CCNF, MAPK14, INCENP, and CHAF1A), which were significantly associated with overall survival (OS) in the test series. Based on this five‐gene signature, the test series patients could be classified into high‐risk or low‐risk subgroups. Multivariate Cox regression analysis indicated that the prognostic power of this five‐gene signature was independent of clinical features. In conclusion, we developed a five‐gene signature related to the cell cycle that can predict survival for GC. Our findings provide novel insight that is useful for understanding cell cycle mechanisms and for identifying patients with GC with poor prognoses.