Identification and Analysis of Key Genes Driving Gastric Cancer Through Bioinformatics.

Objective: The aim of this study was to use bioinformatic analyses to identify key genes and pathways driving gastric cancer (GC). Materials and Methods: The gene expression profiles, from human gastric tissue samples were downloaded from the Gene Expression Omnibus (GSE)29272 dataset. These data revealed 284 differentially expressed genes (DEGs) that included a group upregulated in cancer tissues (n = 142) and another group that were downregulated in cancer tissues. (n = 142). These DEGs were identified using the GEO2R. We used multiple online analysis tools, including, Gene Ontology (GO), Kyoto Encyclopedia of Genes and Genomes (KEGG), protein-protein interaction networks, gene expression profiling interactive analysis (GEPIA), and the cBio Cancer Genomics Portal (cBioportal) database. Next, we identified the most significant DEGs using the Kaplan-Meier plotter (KM-plotter) database. Multiple bioinformatic platforms were used to identify candidate prognostic marker genes. We then analyzed freshly frozen GC tissues for the expression of these marker genes to validate the informatic findings. Results: We identified three DEGs related to overall survival from our analyses of the GEO data. Next, we analyzed these three DEGs in GEPIA and the cBioportal database and found that the biglycan (BGN) gene was related to invasion and metastases of GCs. This finding of differential gene expression was confirmed in a separate laboratory analysis of normal and GC tissues. In this analysis we found that high levels of BGN expression were correlated with GC clinicopathological characteristics, including microvascular tumor thrombus (p = 0.018), lymph node metastases (p = 0.013), and vessel invasion (p = 0.004). Conclusions: BGN expression levels appear to be an independent prognostic factor for predicting the survival times of GC patients.