Identification of novel proteins involved in plant cell-wall synthesis based on protein-protein interaction data

The plant cell wall is mainly composed of polysaccharides, representing the richest source of biomass for future biofuel production. Currently, the majority of the cell-wall synthesis-related (CWSR) proteins are unknown even for model plant Arabidopsis thaliana. We report a computational framework for predicting CWSR proteins based on protein-protein interaction (PPI) data and known CWSR proteins. We predict a protein to be a CWSR protein if it interacts with known CWSR proteins (seeds) with high statistical significance. Using this technique, we predicted 100 candidate CWSR proteins in Arabidopsis thaliana, 8 of which were experimentally confirmed by previous reports. Forty-two candidates have either independent supporting evidence or strong functional relevance to cell-wall synthesis and, hence, are considered as the most reliable predictions. For 33 of the predicted CWSR proteins, we have predicted their detailed functional roles in CWS, based on analyses of their domain architectures, phylogeny, and current functional annotation in conjunction with a literature search. We present the constructed PPIs covering all the known and predicted CWSR proteins at http://csbl.bmb.uga.edu/∼zhouchan/CellWallProtein/. The 42 most reliable candidates provide useful targets to experimentalists for further investigation, and the PPI data constructed in this work provides new information for cell-wall research.