Using a process to which we refer as 'reverse engineering', we propose to reconstruct signaling pathways in benign, aggressive, and metastatic prostate cancer cells from integrated microarray data. Current reverse-engineering signaling pathways have been an ill-posed problem because the signaling pathway topology is difficult, if not impossible, to directly observe in a single experiment, and high experimental costs prohibit collecting sufficient samples in a single study. As a result, current research often relies upon overly stringent, sometimes unrealistic, biological, statistical or computational assumptions. In this proposal, we aim to develop a new informatics paradigm to reconstruct cancer signaling pathways, an approach that will overcome these limitations. The specific hypotheses underlying the proposed research are: 1. Signaling pathways are composed of entangled information flows of overlapping gene sets. 2. The 'Information Flow Gene Set' (IFGS) is the basic functional unit of the signaling pathway, and many IFGS's execute distinct biological endpoint functions. 3. The signaling pathways in benign tissue, in primary cancer tissue and in metastatic cancer tissue are composed of differentially regulated information flows and may be dissected by analysis of the IFGS's. To test these hypotheses we will carry out the following specific aims to reconstruct and compare signaling pathways in different prostate cancer types: SA1. Using integrated gene expression data, we will identify unordered, overlapping and biologically relevant IFGS's for each prostate tissue type. SA2. We will reconstruct realistic signaling pathways from overlapping IFGS's and determine differentially activated/inhibited information flows corresponding to prostate cancer progression.
Among all cancer types, prostate cancer is the most common non-skin cancer. Its etiology remains largely unknown in that more than half of American men have some cancerous cells in their prostate, but most show no signs of disease. Deciphering signaling pathways in benign, primary, and metastatic prostate cancer tissues will help us to understand the mechanisms of carcinogenesis and develop therapeutic approaches based upon the biology of the tumor
