Integrated analysis of gene expression signatures associated with colon cancer from three datasets

Purpose: The present study aimed to elucidate the pathogenesis of colon cancer and identify genes associated with tumor development. Methods: Three datasets, two (GSE74602 and GSE44861) from the Gene Expression Omnibus database and RNA-Seq colon cancer data from The Cancer Genome Atlas data portal, were downloaded. These three datasets were grouped using a meta-analysis approach, and differentially expressed genes (DEGs) were identified between colon tumor samples and adjacent normal samples. Functional enrichment analysis and regulatory factor predication were performed for significant genes. Additionally, small-molecule drugs associated with colon cancer were predicted, and a prognostic risk model was constructed. Results: There were 251 overlapping DEGs (135 up- and 116 downregulated) between cancer samples and control samples in the three datasets. The DEGs were mainly involved in protein transport and apoptotic and neurotrophin signaling pathways. A total of 70 small-molecule drugs were predicated to be associated with colon cancer. Additionally, in the miRNA-target regulatory network, we found that SLC44A1 can be targeted by hsa-miR-183, hsa-miR-206, and hsa-miR-147, while KLF13 can be regulated by hsa-miR-182, hsa-miR-206, and hsa-miR-153. Moreover, the results of the prognostic risk model showed that four genes (VAMP1, P2RX5, CACNB1, and CRY2) could divide the samples into high and low risk groups. Conclusion: SLC44A1 and KLF13 may be involved in tumorigenesis and the metastasis of colon cancer by miRNA regulation. In addition, a four-gene (VAMP1, P2RX5, CACNB1, and CRY2) expression signature may have prognostic and predictive value in colon cancer.