Systems-level modeling of mutationally activated signaling networks and response to therapy
Cancer genomes are characterized by mutations that allow cancer cells to grow in an uncontrolled fashion. The fast growth of cancer cells implies that changes that do not affect cell growth can spread in a tumor lineage as well, so called passenger mutations. A third class of mutations plays a key role once medical treatment commences: resistance mutations. These mutations are modulators that might mildly contribute to tumor growth (or, at least in principle, even not at all), and therefore initially need not be present at high frequency in the tumor. Once treatment starts, however, resistance mutations confer resistance to a specific therapy to cells harboring them. Cells not bearing resistance mutations are affected by therapy, resulting in tumor shrinkage. The cells harboring a resistance mutation survive, keep growing quickly and ultimately lead to relapse and resistance to the initial therapy.
Resistance mutations are (i) overrepresented in tumors which do not respond to therapy, and (ii) they bypass the therapeutic intervention in some fashion at the level of the cancer pathway. We will use both of theses features to identify resistance mutations in genomic data based on a combination of statistical modeling and pathway analysis. This approach is complements the analysis of tumor evolution (SP2), where resistance mutations increase in frequency in the tumor under therapy.
Keywords: cancer genome, statistical analysis