Establishing Protein Functional Linkage in a Systematic Way∗

Abstract Most gene products facilitate their functions within complex interconnected networks by interacting with other biomolecules. Thus elucidating protein functional relationships from their interacting neighbors is one of the challenging problems of the post-genomic era. High-throughput experiments such as genome-wide protein-protein interaction networks are expected to be fertile sources of information for deriving their functional relationships. However, a high rate of false positives and the sheer volume of the data are making reliable interpretation of these experiments difficult. In this work, we overcome these difficulties by using a network-based statistical signif-icance analysis method that forms reliable functional associations between proteins. The basic mechanism is if two proteins share similar neighbors globally than random, they have close func-tional associations. Our method tries to establish a framework to explore the protein relationships by analyzing statistical significance of sharing similar global partnerships for all protein pairs in the interaction network. In this framework, many methods can be integrated to globally define and construct protein neighborhood from protein interaction data. Furthermore our framework can be applied directly to binary data, experimental strength data and integration data. Applying our framework in yeast protein interaction datasets and the shortest path to form protein neighborhood, our method is shown to be able to infer reliable functional linkages from experimental data which are verified by GO functions