Identification modules of gastric cancer based on protein-protein interaction networks and gene expression data.

PURPOSE: The tumorigenesis of gastric cancer is an intricate process which contains genetic and epigenetic changes of proto-oncogenes and cancer-suppressor genes. The purpose of this study was to identify novel modules for gastric cancer based on protein-protein interaction networks and gene expression data. METHODS: Microarray data and corresponding annotated files of E-GEOD-15460 were downloaded from ArrayExpress database. All human protein-protein interactions were downloaded from STRING database. The fast depth-first assay was used to identify all maximal cliques of disease group and control group. Benjamini-Hochberg method was used to perform multiple corrections of p value. RESULTS: 248 modules for the control group and 30 modules for the disease group were determined in this research, and 734 pairs of similar or same modules of these two groups were detected through calculating module correlation density. Protein-protein interaction (PPI) network was identified, which comprised of 7899 genes and 48469 interrelationship pairs of genes. Finally, 6 modules with remarkable difference were found to be closely related with gastric cancer. CONCLUSIONS: Novel modules with significant difference and the related genes are useful biomarkers and therapeutic targets for gastric cancer.