Brain tumors are the most common solid malignancies and the leading cause of cancer- related death in children. Medulloblastoma (MB) is the most common pediatric brain tumor. Dissemination (metastasis) of MB through the cerebrospinal fluid seeds the leptomeningeal membranes that cover the brain and spinal cord. Metastases in MB are refractory to treatment, essentially defining children with incurable tumor. We mobilized the transposable element Sleeping Beauty (SB) in a genetically engineered mouse model of MB driven by Sonic Hedgehog (Shh) signaling and observed robust metastases. Genetic analyses of matched primary and metastatic lesions indicate that tumors undergo parallel evolution and harbor distinct, clonally selected mutations generated by transposition. This is among the first mouse models enabling identification of genes driving metastases. We hypothesize that SB can uncover clonal organization and genes underlying progression/metastasis, and that these data will inform human MB. We propose experiments characterizing paired primary and metastatic tumors from patients with MB. We also propose additional SB based experiments to identify metastases genes in a novel MYCN-driven model for MB which arises largely independently of Shh signaling and which models both classic (60% of human MB) and large cell, anaplastic pathologies (10% of human MB). Because classic human MB can be driven either by myc or by loss of p53, we will also mobilize SB in p53 deficient mice. The use of 3 models minimizes biological effects due to background, characterizes a broad genetic subset of MB, and facilitates identification and prioritization of: 1). Genes driving metastases in 3 distinct models for MB. 2). Candidates metastases genes altered in human MB. 3). Potential therapeutic targets. These data have profound implications for therapy, which assumes that metastases are biologically similar to the primary tumor. A.1.To recover transposon insertion site sequences from matched primary/metastatic GEM MB in two models in order to identify genes and pathways important for MB pathogenesis. A.2.To validate the functional importance of candidate metastases genes identified by SB insertion. A.3.To evaluate hierarchical structures, genes and pathways important for leptomeningeal dispersion using human genomic data, and paired tumors and metastases from human MB.