In this project, we would like to collect and analyze publicly available clinical information and transcriptomic data for patients with different cancers. We will perform Kaplan-Meier and Cox regression survival analysis for all protein coding genes in all cancer types. In addition, we will constructed cancer specific co-expression networks for each of the cancer types, and perform network based analysis to identify key gene clusters and nodes which might be potential therapeutic targets. Moreover, we will reconstruct a comprehensive metabolite-gene network by integrating information from KEGG, Reactome and STRING database, and use reporter metabolite analysis to identify hub metabolites that are potentially important to survival outcome of cancer patients. Finally, metabolic subnetworks will be built by connecting hub metabolites identified for each cancer to reveal the key metabolic pathways that is important for cancer patient survival.