Comprehensive Analysis Of Predicted Rice Protein-protein Interaction Network

Computational identification of protein-protein interaction (PPI) network is now a hotspot in biological science research community. Here we applied bioinformatics approaches to infer potential PPIs for Oryza sativa by using interolog information. In order to deeply understand the rice interactome, We analyzed GO annotation, subcellular localization, and gene expression of 5,049 rice proteins with 76,585 interactions.84% proteins in the predicted network were highly annotated by Gene Ontology (GO), showing significant tendencies in co-GO-annotation. After gene expression analysis, we found interacted proteins show significant tendencies in co-expression. We also obtained subcellular localization information of 14,308 interactions, in which 49.1% is co-localized. Further analysis of network topology properties shows that the network has significant scale-free property and the small world property. The high modularization of rice interactome was dicovered after comparing with yeast, human and Arabidopsis thaliana. Through network function analysis, top degree proteins and most conserved interaction are derived for further biological function analysis. Additionally, we took MADS-box domain containing proteins and Circadian rhythm signaling pathways as examples to present that proteins functional complex and biological pathways could be effectively expanded in the predicted network.