Non-Hodgkin lymphomas (NHL) represent the sixth most commonly diagnosed type of cancer in Canada with diffuse large B-cell lymphoma (DLBCL) being the most prevalent. DLBCL patients exhibit variable responses to available therapies such as R-CHOP. Some biomarkers beginning to enter clinical trials rely on gene expression signatures to identify high-risk cases. For example, DLBCL comprises two cell-of-origin (COO) subgroups with activated B-cell (ABC) cases typically having worse prognosis than germinal centre B-cell (GCB) cases. Therapeutics that exploit molecular features unique to each subgroup are being actively pursued but up-front identification of patients unlikely to achieve a durable response remains a challenge. Although existing methods can identify specific prognostic gene expression features, these capture only some of the clinical heterogeneity and require tumour RNA, which is not always available. Using comprehensive genomic analyses and by integrating mutation and expression data from over 2000 DLBCL tumours, we seek novel molecular features that are associated with treatment resistance in DLBCL. The ultimate goal of this work is to develop new predictive and prognostic biomarkers and identify new therapeutic targets for treatment-resistant DLBCLs.