Over the past five years, the landscape of treatment for metastatic melanoma has radically changed with the approval of multiple therapies, both targeted and immunotherapies. However, even with these vast improvements in therapeutic options, patients have tumors that demonstrate either intrinsic resistance or develop acquired resistance. It is essential to identify the determinants of resistance. Utilizing specimens from several targeted therapy clinical trials, utilizing BRAF, MEK and CRAF inhibitors, we have identified genetic and genomic determents of progression free and overall survival. Not surprisingly, the predictors fall within the MAPK and PI3K/AKT signaling pathways, and in many cases are linked to target of the therapy. In one such example, in E2603, a randomized phase III comparing carboplatin, paclitaxel, +/- sorafenib (CP vs. CPS) RAF1 (cRAF) amplification is associated with improved PFS in response to treatment with CPS compared to CP (HR=0.372, P=0.025). One of the conundrums of evaluating biomarkers in this context is determining whether they are prognostic or predictive. In many cases, it may that markers currently touted as predictive of response to MAPK inhibitors are prognostic, and have been incompletely evaluated in studies prior to the era of targeted therapy. In E2603, we found that BRAF copy gain was associated with worse progression free and overall survival (P=0.0233 and P=0.0457, respectively), independent of somatic mutation status. BRAF amplification has been suggested to be a predictive marker by many, but is clearly prognostic in this study done prior to the advent of mutant BRAF targeted therapy. We also have used targeted massively parallel sequencing to profile over 300 cell lines and patient derived xenografts so that response to multiple different targeted therapies can be modeled in vivo alone and in combination. With this large sample set, we also are able to identify previously undescribed sub-groups of melanoma.