Bioinformatics analysis of survival-related genes in pancreatic cancer based on GEO and TCGA database / 西安交通大学学报(医学版)

ABSTRACT

【Objective】 Based on Gene Expression Omnibus (GEO) database and The Cancer Genome Atlas (TCGA) database, survival analysis was used to screen the key prognostic genes involved of pancreatic cancer patients. 【Methods】 Two pancreatic cancer gene chips (Microarray) from the GEO database and transcriptome sequencing (RNA-seq) from the TCGA database were used to filter the survival-related genes using Kaplan-Meier (KM) analysis and Cox risk model, and the target genes were intersected. Prognosis-associated genes were screened first and then pathway enrichment analysis or immune-enrichment analysis was performed based on these genes to find out their potential molecular mechanisms in regulating pancreatic cancer. 【Results】 In this study, five survival-related genes (i.e., CDO1, DCBLD2, FAM83A, ITGA3 and SLC16A3) were screened out. Multifactorial Cox regression analysis and clinical correlation analysis showed that high CDO1 expression was a protective factor for pancreatic cancer prognosis, and its antitumor effect was associated with its role in inhibiting the malignant biological behavior of pancreatic cancer cells and promoting the infiltration of immune killer cells in pancreatic cancer. 【Conclusion】 This study suggests that CDO1 is a potential tumor suppress gene of pancreatic cancer, and the tumor inhibition effect of CDO1 may be related to its role in remodeling the immune microenvironment of pancreatic cancer.