A system level analysis of gastric cancer across tumor stages with RNA-seq data.

Gastric cancer is the third leading cause of cancer-related death in the world. Over the past few decades, with the development of high-throughput technologies and the application of various statistical tools, cancer research has witnessed remarkable advancements. However, no system level analysis has taken into account the cancer stages, which are known to be extremely important in prognosis and therapy. In this study, we aimed to carry out a system level analysis of the dynamics of the network structure across the normal phenotype and the four tumor stage phenotypes. We analyzed 276 samples of primary tumor tissues including normal and four tumor stage phenotypes to reveal the dynamics of the five phenotype-specific co-expression networks. Our analysis reveals that the structure of the normal network is dramatically different from that of a tumor network. The analysis of connectivity dynamics shows that hub genes present in the normal network but not in the tumor networks play important roles in tumorigenesis and hub genes unique to a tumor network are enriched in specific biological terms. Moreover, we found three interesting clusters of genes which possess specific dynamic features across the five phenotypes and are enriched in stage-specific biological terms. Integrating the results from the expression analysis and the connectivity analysis shows that the stages of tumor should be taken into consideration and a system level analysis serves as a complement to and a refinement of the traditional expression analysis.