The goal of this research is to develop methods and computational infrastructure for launching two major types of queries on the interaction network of cellular signaling and regulatory pathways. These include: (1) queries to identify network regions that are conserved across multiple species, conditions, or points in time, and (b) queries to match observed genetic interactions, such as synthetic lethals, suppressor mutations, and complex genetic disease linkages, with the underlying pathways of physical interactions that best explain them. Research for query (1) will build on the PathBLAST network comparison algorithm by adding support for protein domains and multiple network alignment, automated visualization and layout of aligned network regions, and a general model of protein insertions and deletions in the aligned networks. Proposed research for query (2) is more exploratory and will evaluate and refine an initial formulation of the problem as a probabilistic causal model. Successful strategies for integrating molecular interactions with other genomic data streams will be crucial for building large-scale wiring diagrams of the cell in a top-down and hypothesis-driven fashion
