Identification of intestinal flora-related key genes and therapeutic drugs in colorectal cancer

• Main Authors: Jiayu Zhang, Huaiyu Zhang, Faping Li, Zheyu Song, Yezhou Li, Tiancheng Zhao
• Format: Article
• Language: English
• Published: BMC 2020-11-01
• Series: BMC Medical Genomics
• Subjects: Colorectal cancer; Intestinal flora; Text mining; Key genes; Drugs
• Online Access: http://link.springer.com/article/10.1186/s12920-020-00810-0

Abstract Background Colorectal cancer (CRC) is a multifactorial tumor and a leading cause of cancer-specific deaths worldwide. Recent research has shown that the alteration of intestinal flora contributes to the development of CRC. However, the molecular mechanism by which intestinal flora influences the pathogenesis of CRC remains unclear. This study aims to explore the key genes underlying the effect of intestinal flora on CRC and therapeutic drugs for CRC. Methods Intestinal flora-related genes were determined using text mining. Based on The Cancer Genome Atlas database, differentially expressed genes (DEGs) between CRC and normal samples were identified with the limma package of the R software. Then, the intersection of the two gene sets was selected for enrichment analyses using the tool Database for Annotation, Visualization and Integrated Discovery. Protein interaction network analysis was performed for identifying the key genes using STRING and Cytoscape. The correlation of the key genes with overall survival of CRC patients was analyzed. Finally, the key genes were queried against the Drug-Gene Interaction database to find drug candidates for treating CRC. Results 518 genes associated with intestinal flora were determined by text mining. Based on The Cancer Genome Atlas database, we identified 48 DEGs associated with intestinal flora, including 25 up-regulated and 23 down-regulated DEGs in CRC. The enrichment analyses indicated that the selected genes were mainly involved in cell–cell signaling, immune response, cytokine-cytokine receptor interaction, and JAK-STAT signaling pathway. The protein–protein interaction network was constructed with 13 nodes and 35 edges. Moreover, 8 genes in the significant cluster were considered as the key genes and chemokine (C-X-C motif) ligand 8 (CXCL8) correlated positively with the overall survival of CRC patients. Finally, a total of 24 drugs were predicted as possible drugs for CRC treatment using the Drug-Gene Interaction database. Conclusions These findings of this study may provide new insights into CRC pathogenesis and treatments. The prediction of drug-gene interaction is of great practical significance for exploring new drugs or novel targets for existing drugs.