Identifying the critical components of cancer networks
We know that large networks of interacting genes and proteins support the functioning of cells in the human body. But we don’t have a precise model for how these networks combine the information in the genetic code to determine traits like height or disease susceptibility. Using cancer as an example, I will present two new ways to analyze the topology of biological networks and identify critical features that functionally affect the phenotype. First, I will show how we can find cancer driver mutations by combining the degree centrality of nodes in the transcriptional network and the protein interaction network to identify bottlenecks of information flow. Second, I will describe an algorithm for detecting the biggest changes in community structure between networks active in two different conditions, like healthy and diseased tissue. When applied to transcriptional networks in ovarian cancer, for example, this method allows us to detect both known and novel pathways associated with poor prognosis. Ultimately, our aim is to build a suite of mathematical methods for predicting how genetic variants perturb biological networks and collectively alter complex phenotypes.