Multiple myeloma (MM) is an incurable malignancy of bone marrow plasma cells characterized by wide clinical and molecular heterogeneity. In this study we applied an integrative network biology approach to molecular and clinical data measured from 450 patients with newly diagnosed MM from the MMRF (Multiple Myeloma Research Foundation) CoMMpass study. A novel network model of myeloma (MMNet) was constructed, revealing complex molecular disease patterns and novel associations between clinical traits and genomic markers. Genomic alterations and groups of coexpressed genes correlate with disease stage, tumor clonality and early progression. We validated CDC42BPA and CLEC11A as novel regulators and candidate therapeutic targets of MMSET-related myeloma. We then used MMNet to discover novel genes associated with high-risk myeloma and identified a novel four-gene prognostic signature. We identified new patient classes defined by network features and enriched for clinically relevant genetic events, pathways and deregulated genes. Finally, we demonstrated the ability of deep sequencing techniques to detect relevant structural rearrangements, providing evidence that encourages wider use of such technologies in clinical practice. An integrative network analysis of CoMMpass data identified new insights into multiple myeloma disease biology and provided improved molecular features for diagnosing and stratifying patients, as well as additional molecular targets for therapeutic alternatives.