Metastasis is the result of stochastic genomic and epigenetic events leading to gene expression profiles that drive tumor dissemination. Here we exploit the principle that metastatic propensity is modified by the genetic background to generate prognostic gene expression signatures that illuminate regulators of metastasis. We also identify multiple microRNAs whose germline variation is causally linked to tumor progression and metastasis. We employ network analysis of global gene expression profiles in tumors derived from a panel of recombinant inbred mice to identify a network of co-expressed genes centered on Cnot2 that predicts metastasis-free survival. Modulating Cnot2 expression changes tumor cell metastatic potential in vivo, supporting a functional role for Cnot2 in metastasis. Small RNA sequencing of the same tumor set revealed a negative correlation between expression of the Mir216/217 cluster and tumor progression. Expression quantitative trait locus analysis (eQTL) identified cis-eQTLs at the Mir216/217 locus, indicating that differences in expression may be inherited. Ectopic expression of Mir216/217 in tumor cells suppressed metastasis in vivo. Finally, small RNA sequencing and mRNA expression profiling data were integrated to reveal that miR-3470a/b target a high proportion of network transcripts. In vivo analysis of Mir3470a/b demonstrated that both promote metastasis. Moreover, Mir3470b is a likely regulator of the Cnot2 network as its overexpression down-regulated expression of network hub genes and enhanced metastasis in vivo, phenocopying Cnot2 knockdown. The resulting data from this strategy identify Cnot2 as a novel regulator of metastasis and demonstrate the power of our systems-level approach in identifying modifiers of metastasis.[Supplemental material is available for this article.]Metastasis is a systemic disease responsible for the majority of cancer-related mortality and is influenced by both tumor cell-autonomous and host-derived factors. Its complexity is deepened by involvement of not only stochastic genomic and epigenetic events but also by inherited predisposition (Lifsted et al. 1998;Crawford et al. 2006). As a result, despite identification and characterization of individual genes, cellular and developmental processes associated with metastasis, understanding of the metastatic cascade and the interconnectivity of individual factors remains limited. The elucidation of higher-order networks underlying metastasis will therefore likely improve prognostication and intervention strategies by identifying molecular nodes central to tumor cell dissemination and colonization.Recent advances in global gene expression profiling and computational science have provided the basis for understanding cancer biology at a systems level (Quigley et al. 2009). Knowledge of both tumor subtypes (Perou et al. 2000) and patient prognosis has been enhanced by systems-based approaches. This knowledge may significantly change clinical practice by enabling the development of precision treatments based on mol...