| I have combined four different kinds of functional genomic data to create high coverage, probabilistic protein interaction networks for 312 microbes. My integration algorithm naturally handles statistically dependent predictors and automatically corrects for differing noise levels and data corruption in different evidence sources. I find that a plurality of the predictions in each integrated network hinge on moderate but consistent evidence from multiple sources rather than strong evidence from a single source, yielding novel biology that would be missed if a single data source such as coexpression or coinheritance were used in isolation.; I demonstrate that these hidden interactions uncover new aspects of well studied functional modules in a broad range of microbial species. I accompany this analysis with a strategy for systematic, computer-guided laboratory validation. As this work represents the largest collection of probabilistic protein interaction networks compiled to date, I have worked with collaborators to create tools for network alignment to organize this information. By developing the first scalable multiple network alignment algorithm, we have identified thousands of conserved modules in diverse microbial species. Our integration, validation, and alignment algorithms can be applied to any sequenced organism and any kind of experimental or computational technique which produces pairwise measures of protein interaction. |
