Identification of sixteen metabolic genes as potential biomarkers for colon adenocarcinoma.

PURPOSE To identify some key prognosis-related metabolic genes (PRMG) and establish a clinical prognosis model for colon adenocarcinoma (COAD) patients. METHODS We used The Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GEO) to obtain gene expression profiles of COAD, and then identified differentially expressed prognostic-related metabolic genes through R language and Perl software, Through univariate Cox analysis and least absolute shrinkage and selection operator (LASSO) Cox analysis to obtain target genes, established metabolic genes prognostic models and risk scores. Through Cox regression analysis, independent risk factors affecting the prognosis of COAD were analyzed, and receiver operating characteristics (ROC) curve analysis of independent prognostic factors was performed and a nomogram for predicting overall survival was constructed. We performed the consistency index (C-index) test and decision curve analysis (DCA) on the nomogram, and used gene set enrichment analysis (GSEA) to identify the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway of model genes. We selected PRMG based on the expression of metabolic genes, and used LASSO Cox regression to construct 16 metabolic gene models (SEPHS1, P4HA1, ENPP2, PTGDS, GPX3, CP, ASPA, POLR3A, PKM, POLR2D , XDH, EPHX2, ADH1B, HMGCL, GPD1L and MAOA). RESULTS The risk score generated from our model can well predict the survival prognosis of COAD. A nomogram based on the clinicopathological characteristics and risk scores of COAD can personally predict the overall survival rate of COAD patients. CONCLUSIONS The risk score based on the expression of 16 metabolic genes can effectively predict the prognosis of patients with COAD.