| Esophageal squamous cell carcinoma (ESCC) is one of the most common cancers in China with5-year overall survival rate less than25%. The high mortality and poor prognosis of ESCC patients is usually ascribed to the relapse and the metastasis. And the occurrence and development of ESCC is a complex biological process which involves the interaction between multiple genes, transcripts and proteins. Thus, comprehensive research on the molecular mechanisms would improve the effect of diagnosis and treatment of ESCC.Traditional biology research can reveal the mechanism of life at the molecular level based on the variation and function of individual gene, mRNA, or protein, but it only gives partial explanation to biological phenomenon instead of describing it in the whole system. However, biological networks provides a novel platform to study the characteristics of biology at the system level and elaborate the relationship of different functional elements.The general framework for weighted gene co-expression analysis (WGCNA) is based on a systems biology method for describing the correlation patterns among genes and finding modules of highly correlated genes across microarray samples. It considers not only the co-expression patterns between two genes but also the overlap of neighbouring genes. This study used WGCNA to establish a human ESCC gene co-expression network from published microarray gene expression data. We chose the differential expressing genes between tumors tissues and adjacent normal tissues as the data input for WGCNA, and we have identified8gene modules related to epidermal cell differentiation, extracellular matrix organization, cell cycle process and so on. Then an independent cohort of ESCC expression profiles was chosen for consensus module detection between the original and independent cohort. We found that most specific modules from the original cohort shared consensus counterparts with the independent cohort which meant that the module structure in the original cohort expression data was similar to the independent cohort data. Finally, we used combined analysis and identified several gene modules closely related to the clinical information including tumor grading, patient survival time, and tumor T staging. In addition, we confirmed that several hub genes with high levels of connectivity were highly correlated with tumor grading. These genes may potentially play an important role in ESCC.In conclusion, this was the first time that WGCNA was applied to the study of ESCC. Our results demonstrated that WGCNA is able to identify gene modules with biological meanings. Moreover, the hub genes we found were consistent with the previous reports, proving the accuracy and effectiveness of this algorithm. In the future, more additional research should be done to understand the regulatory mechanisms of critical genes, signaling pathways, and the interactions between genes of ESCC. |
