Pan-Cancer Proteogenomic Network Analysis Uncovers Immune Microenvironment and Prognostic Biomarkers

Abstract

Pan-cancer analysis provides valuable insights into common and distinct molecular mechanisms across various cancer types. Co-expression networks, which captured the cordinated expression of genes and proteins, offer potentials for identifying robust biomarkers and targeted therapies. However, current-existing co-expression construction method involves either RNA or Protein data in isolation, limiting our understanding on multi-perspective interactions in cancer. To address this gap, we integerated both the RNA and Protein data to constrcut a comprehensive pan-cancer co-expression network, combining both gene and proteomic layers of information across cancer types. This approach enables the identification of previously undetected co-expression relationships, highlighting key regulatory patterns in cancer analysis. Our study constructed a multi-perspective co-expression network, after that, we identified several conservation modules and compare the modules identified from RNA and Protein. We found out that the conservation modules identified differ significantly from RNA and Protein data. However, there are certain shared conservation modules, especially those related to IFN-gamma, IFN-alpha, and the activation of JAK_STAT. Then we identified several important immune-related pathways using RNA and Protein data and uncovered some modules with high prognostic value across cancers. As a novel pan-cancer analysis approach, it holds promise for advancing future cancer research, from the discovery of novel biomarkers to the design of personalized cancer therapies, thereby enabling more thorough and holistic cancer analysis.