PT - JOURNAL ARTICLE AU - Jean Hausser AU - Pablo Szekely AU - Noam Bar AU - Anat Zimmer AU - Hila Sheftel AU - Carlos Caldas AU - Uri Alon TI - Universal cancer tasks, evolutionary tradeoffs, and the functions of driver mutations AID - 10.1101/382291 DP - 2018 Jan 01 TA - bioRxiv PG - 382291 4099 - http://biorxiv.org/content/early/2018/08/01/382291.short 4100 - http://biorxiv.org/content/early/2018/08/01/382291.full AB - Recent advances have led to an appreciation of the vast molecular diversity of cancer. Detailed data has enabled powerful methods to sort tumors into groups with benefits for prognosis and treatment. We are still missing, however, a general theoretical framework to understand the diversity of tumor gene-expression and mutations. To address this, we present a framework based on multi-task evolution theory, using the fact that tumors evolve in the body, and that tumors are faced with multiple tasks that contribute to their fitness. In accordance with the theory, we find that tradeoff between tasks constrains tumor gene-expression to a continuum bounded by a polyhedron. The vertices of the polyhedron are gene-expression profiles each specializing in one task, allowing the tasks to be identified. We find five universal cancer tasks across tissue-types: cell-division, biomass & energy, lipogenesis, immune-interaction and invasion & tissue remodeling. Tumors whose gene-expression lies close to a vertex are task specialists. We find evidence that such specialists are more sensitive to drugs that interfere with this task. We find that driver mutations, but not passenger mutations, tune gene-expression towards specialization in specific tasks. This approach can integrate additional types of molecular data into a theoretically-based framework for understanding tumor diversity.