Improved computational prediction and analysis of protein - protein interaction networks

Abstract

Recent progress in computational methods for predicting physical and functional protein-protein interactions has provided new insights into the complexity of biological processes. Most of these methods assume that functionally interacting proteins are likely to have a shared evolutionary history. These methods include phylogenetic profiling (PP),gene neighborhood (GN), gene cluster (GC), and the mirrortree. Expression similarity in various physiological conditions also has been used an indicator of functional linkages between genes. Comprehensive newlinecomparison of these methods has not been frequently reported in literature. In this work, I have shown that the higher performance for predicting protein-protein interactions was achievable even with 100?150 bacterial genomes out of 565 genomes. I find that variants of PP and GN are robust against reference genome selection. This study also reveals that the prediction of metabolic pathway protein interactions continues to be a challenging task for all methods which possibly reflect flexible/independent evolutionary histories of these proteins. On the contrary, genetic information processing pathways are predicted with comparable accuracy. I have also shown that the effective use of a particular prediction method depends on the pathway under investigation. The topological properties of network predicted by each method differ significantly. This study suggests that organization of proteins in the predicted networks ensure the local perturbations in the metabolic pathways and protein complexes should communicate with quickly to other cellular proteins. A set of seven machine learning classifiers also used to predict genome-scale interactome. It is observed that probabilistic classifiers such as naïve bayes are best suitable for PPI prediction task. Finally, I have predicted functions for number of uncharacterized proteins and some of them tested experimentally. A high quality PPIs can be accessible through user friendly interface at http://www.cdfd.org.in/ecofunppi