Complexity of human cancer is driven by the coordinated activation and inactivation of multiple genes, which makes the identification of causal drivers of cancer progression a daunting challenge. Although animal models are often used to study mechanisms of cancer progression and evaluate new cancer therapies, the accurate extrapolation of animal studies to human cancer has been difficult. I will present novel cross-species systems biology algorithms that identify conserved regulatory programs between human and mouse cancer models and inform on therapeutic strategies for human patients with the most aggressive disease. These algorithms identify causal gene “drivers” of aggressive cancer, which may also serve as biomarkers to categorize patients with poor prognosis. We have generated complementary human and mouse prostate cancer gene regulatory networks (interactomes) assembled from molecular profiles of human tumors and genetically engineered mouse models. Our computational systems biology network-based approaches and subsequent experimental validation have elucidated a synergistic interaction of two genes, FOXM1 and CENPF, that drives prostate cancer aggressiveness and is a robust prognostic indicator of cancer outcome. I will demonstrate that these identified drivers are excellent candidates for targeted therapeutics, especially for patients with aggressive prostate cancer. Furthermore, I will describe an innovative computational algorithm to identify drugs and drug combinations that inhibit the transcriptional activity of these molecular drivers. Experimental validation confirms high efficacy of the top predicted drug combination for inhibiting tumorigenesis in mouse and human prostate cancer models. Although these approaches have been specifically applied to prostate cancer, they also address issues of broad general relevance for the prognosis, diagnosis, and treatment of human disease.