Screening and identification of potential biomarkers for pancreatic cancer: An integrated bioinformatics analysis

Abstract

Pancreatic cancer is one of the highly invasive and the seventh most common cause of death among cancers worldwide. To identify essential genes and the involved mechanisms in pancreatic cancer, we used bioinformatics analysis to identify potential biomarkers for pancreatic cancer management. Gene expression profiles of pancreatic cancer patients and normal tissues were screened and downloaded from The Cancer Genome Atlas (TCGA) bioinformatics database. The Differentially expressed genes (DEGs) were identified among gene expression signatures of normal and pancreatic cancer, using R software. Then, enrichment analysis of the DEGs, including Gene Ontology (GO) analysis and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis, was performed by an interactive and collaborative HTML5 gene list enrichment analysis tool (enrichr) and ToppGene. The protein-protein interaction (PPI) network was also constructed using the Search Tool for the Retrieval of Interacting Genes (STRING) database and ToppGenet web based tool followed by identifying hub genes of the top 100 DEGs in pancreatic cancer using Cytoscape software. Over 2000 DEGs with variable log2 fold (LFC) were identified among 34,706 genes. Principal component analysis showed that the top 20 DEGs, including H1–4, H1–5, H4C3, H4C2, RN7SL2, RN7SL3, RN7SL4P, RN7SKP80, SCARNA12, SCARNA10, SCARNA5, SCARNA7, SCARNA6, SCARNA21, SCARNA9, SCARNA13, SNORA73B, SNORA53, SNORA54 might distinguish pancreatic cancer from normal tissue. GO analysis showed that the top DEGs have more enriched in the negative regulation of gene silencing, negative regulation of chromatin organization, negative regulation of chromatin silencing, nucleosome positioning, regulation of chromatin silencing, and nucleosomal DNA binding. KEGG analysis identified an association between pancreatic cancer and systemic lupus erythematosus, alcoholism, neutrophil extracellular trap formation, and viral carcinogenesis. In PPI network analysis, we found that the different types of histone-encoding genes are involved as hub genes in the carcinogenesis of pancreatic cancer. In conclusion, our bioinformatics analysis identified genes that were significantly related to the prognosis of pancreatic cancer patients. These genes and pathways could serve as new potential prognostic markers and be used to develop treatments for pancreatic cancer patients.