INTEGRATED COMPUTATIONAL NETWORK ANALYSIS IDENTIFIES CANDIDATE BIOMARKERS IN COLORECTAL CANCER

Abstract

Colorectal cancer (CRC) is one of the leading causes of cancer morbidity and mortality and the molecular biomarkers may be useful for early detection and biological stratification. In this study, a comprehensive bioinformatics pipeline was employed to determine differentially expressed genes (DEGs), hub protein interactions, enriched functions and potential microRNA/transcription factor regulatory networks in CRC. Gene-expression data sets for the CRC were obtained from the publicly available NCBI Gene Expression Omnibus (GEO) database. The study thresholds were p value < 0.05 and log2 fold change > 2 (|log2 fold change| >= 2) for differential expression analysis using GEO2R/Limma. The overlapping DEGs were analyzed with STRING for protein-protein interaction (PPI) and Cytoscape/cytoHubba for network centrality. The functional enrichment was performed using DAVID for Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways. Functional and mRNA-miRNA-transcription factor relationships were explored using GeneMANIA and miRNet, whereas GEPIA was used for evaluating the expression of tumor versus normal. Three GEO datasets were kept for analysis. GSE130750 resulted in 63 upregulated genes and 433 downregulated genes, while GSE160988 resulted in 134 upregulated and 606 downregulated genes. The third data set provided a smaller set of differential genes, 16 of them down and 0 up, reported in the thesis. A list of highly-connected hub genes was obtained via network analysis. The genes that were reported were extracellular-matrix/collagen: COL4A1, COL4A2, COL5A1, COL5A2, COL6A1, COL6A3, COL12A1 and immune-inflammatory: CXCL10, CCL18, IDO1, ITGAX, NLRP3, P4HA3 and TNFRSF1B. The regulatory-network analysis linked hub genes to the candidate miRNAs and transcription factors. The integrated analysis provides valuable insight into candidate CRC-associated genes and regulatory relationships that could be targeted for further validation as diagnostic, prognostic or therapeutic biomarkers. Before clinical interpretation, independent cohorts, correction for multiple testing and laboratory/clinical validation are required.