Increasing the reliability of protein interactomes.

Protein interactions are crucial components of all cellular processes. An in-depth knowledge of the full complement of protein interactions in a cell, therefore, provides insight into the structure, properties and functions of the cell and its components. An accurate and comprehensive protein interaction network is, thus, an invaluable framework to study protein regulation in disease. Although the amount of protein-protein interaction data has grown significantly because of advances in high-throughput experimental techniques, these high-throughput methods are highly susceptible to noise. Therefore, computational techniques for assessing the reliability of a protein-protein interaction are highly desirable. We review here computational techniques for assessing and improving the reliability of protein-protein interaction data from these high-throughput experiments.