Copy number variation, an important kind of genomic variation, has gained increasing attention in recent years mainly due to SNP microarray technology which has made studying whole genome fast and economical. The importance of CNVs to occurrence and development of disease has been confirmed in many studies. Until now, most studies of CNVs are focused on CNVs�� impact on expression of genes located in verified regions, like eQTL, a linear-regression based method. Others may combine CNV with network method, like co-expression network to analyze CNVs�� impact on not just genes inside CNV regions but also outside CNV regions that are co-expressed. But there is little work about interpreting influence of genomic variation on expression through its disturbance to TRN. Mutation in TFs can cause huge cascade effects as a TF targets a large amount of genes involving many biological processes. For example, TP53, a well-known tumor suppressor PS 1145 dihydrochloride transcription factor, its mutation has been reported associated with cell migration and invasion. In 2012, David et al detailed three mutated transcriptional factors NKX2-5, GATA4, and TBX5 and their affected Ro 01-6128 pathways in congenital heart disease. Essaghir et al introduced an integrated approach to construct minimal connected network to TFs in 305 different human cancer cell lines and found several universal cancer biomarkers. These researches suggest the importance and feasibility of integrating TRN with CNVs. Intrahepatic cholangiocarcinoma is the second most common primary hepatic cancer with the highest occurring rate in Thailand and other eastern Asian areas due to chronic inflammation of bile ducts. In 2013, Sia et al performed gene expression and copy number variation integrated analysis in ICC samples and classified these samples into two groups: proliferation and inflammation. Pathogenesis studies based on gene expression profiling have evolved through several stages: single gene expression profiling; network construction and functional annotation; causal hub discovery and intervention design. Single gene expression profiling is straightforward and simple, numerous gene list signatures have been reported to either diagnose samples or predict outcome or prognosis. However it is hard to explain the functional categories of single genes. Network analysis allows structured grouping of genes, and functional module discovery can often lead to next-step research focus, which is a big progress compared to single gene profiling.