Analysis and application of RNA binding protein gene pairs to predict the prognosis of gastric cancer

BackgroundRNA-binding proteins (RBPs) are closely related to tumors, but little is known about the mechanism of RBPs in tumorigenesis and progression gastric cancer (GC). As genes do not usually act alone pathway deregulation, gene pair combinations more likely become stable accurate biomarkers. The purpose our research establish a novel signature based on RBP pairs predict prognosis patients.MethodsWe downloaded genetic clinical information from TCGA GEO database. GSE13911 were used for screening differentially expressed (DEGs). gathered previous studies employed screen out DE-RBP after intersecting with DEGs. Samples classified according relative expression each genes. univariate Cox regression analysis random forest identify hub construct predicting cancer. Time-dependent ROC curves KM survival performed evaluate signature. GSEA was training cohort GSE62254 testing analyze enrichment pathways. Finally, influence these GC patients further elucidated respectively through combination high low two pair.ResultsWe screened 6 (COL5A2/FEN1, POP1/GFRA1, EXO1/PLEKHS1, SLC39A10/CHI3L1, MMP7/PPP1R1 B SLC5A6/BYSL) Using optimal cut-off value divide into high-risk low-risk groups cohort, we found that overall (OS) group higher than (P < 0.05). area under 1, 3, 5 years (0.659, 0.744, 0.758) (0.624, 0.650, 0.653) cohorts. Univariate multivariate showed independent prognostic factors In addition, COL5A2-high/FEN1-low, POP1-low/GFRA1-high, EXO1-low/PLEKHS1-low,SLC39A10-high/CHI3L1-low, MMP7-high/PPP1R1 B-low, SLC5A6-low/BYSL-low had worse OS And correlation there no obvious between except SLC5A6/BYSL POP1/GFRA1. enriched tumor migration, invasion growth-related pathways.ConclusionOur study identified provide potential targets therapy.