Glioblastomas (GBM) are the most common and aggressive type of adult brain tumors. Despite the standard of care, concomitant radiation and temozolomide based chemotherapy treatment, disease relapse typically occurs within a year for most patients. The complexity of the disease is underlined by recent discoveries of regional differences within the tumor that may contribute to therapy resistance. Through high throughput sequencing and tagging of individual GBM cells using shRNA barcodes, the heterogeneity of the tumor cell mix can be investigated under variable circumstances. The first goal of this grant is to apply this methodology when growing GBMs in mice, to evaluate the degree of complexity after proliferation in absence of therapeutic challenges. To study whether the heterogeneity of the tumor cell mix contributes to the sensitivity to chemo- and radio-therapy, we will apply standard treatment protocols to mouse xenografts and analyze the cellular diversity of the resulting tumors. The second aim of this grant is to evaluate whether tumor complexity can be modulated using therapeutics and whether this property plays a role in developing treatment resistance. Through computational and mathematical approaches, the genomic abnormality profile can be analyzed to infer clonal and subclonal cell populations. When applied to multiple related genomic profiles, such as from diagnostic tumors and matching post-treatment tumor biopsies, patterns of clonal evolution can be uncovered. These can be related to patient features such as outcome, but also to tumor biology characteristics such as the presence of specific genomic alterations. The final aim of this grant is to construct the evolutionary path tha GBM take to escape treatment and result in recurrence. In summary, by evaluating the patterns of clonal evolution of single cells under normal growth properties, under the stress of treatment and in patient tumors, this proposal aims to improve our understanding of why GBM are so resistant to the toxic effects of therapy.