Genetic variations underlie multiple phenotypes, both at the cellular and at the clinical level. However, the complexity of most traits and the burden of multiple hypotheses make it difficult to uncover the genetic mechanisms that cause phenotypic diversity. In this talk, I will describe computational methods for using biological networks to model and understand the effect of genetic variations. We show how gene expression data from a population of genetically diverse individuals (eQTL data) can be used both to uncover regulatory networks and to understand the mechanisms by which genetic variations perturbs those networks. We also show how a prior knowledge and data regarding biological networks and mechanisms can be used to guide the construction of regulatory networks and allow the robust selection of genetic variations that are causal for phenotypic change. We demonstrate the applicability of these methods to problems ranging from understanding mRNA degradation pathways to elucidating mechanisms involved in human disease.
