Gene Co-Expression Network Analysis of Two Ovarian Cancer Datasets

Shengjun Hong,Hua Dong,Li Jin,Momiao Xiong

DOI: https://doi.org/10.1109/bibmw.2010.5703811

2010-01-01

Abstract:Ovarian cancer is one of the leading causes of death in women. To describe the complex gene regulatory relationships and investigate genes acting important roles in ovarian cancer, we adopted gaussian graphic model to construct gene co-expression networks of two independent ovarian cancer datasets separately. To validate the robustness of networks, modules are identified by decision tree cut algorithm and their functions were investigated. Our results showed that the inferred networks were structurally conservative and the identified modules were highly overlapped across the datasets. We discovered four conserved modules which were enriched with the genes in four cancer related pathways. Besides, we detected an ovarian cancer related gene CCEN2 and other six cancer related genes which may also play important roles in ovarian cancer. All the above results showed that incorporating gene co-expression network into the gene expression analysis may facilitate the discovery of cancer mechanisms.