In silico approach to predict candidate R proteins and to define their domain architecture.

Background: Plant resistance genes, which encode R-proteins, constitute one of the most important and widely investigated gene families. Thanks to the use of both genetic and molecular approaches, more than 100 R genes have been cloned so far. Analysis of resistance proteins and investigation of domain properties may afford insights into their role and function. Moreover, genomic experiments and availability of high-throughput sequence data are very useful for discovering new R genes and establish hypotheses about R-genes architecture.

Result: We surveyed the PRGdb dataset to provide valuable information about hidden R-protein features. Through an in silico approach 4409 putative R-proteins belonging to 33 plant organisms were analysed for domain associations frequency. The proteins showed common domain associations as well as previously unknown classes. Interestingly, the number of proteins falling into each class was found inversely related to domain arrangement complexity. Out of 31 possible theoretical domain combinations, only 22 were found. Proteins retrieved were filtered to highlight, through the visualization of a Venn diagram, candidate classes able to exert resistance function. Detailed analyses performed on conserved profiles of those strong putative R proteins revealed interesting domain features. Finally, several atypical domain associations were identified.

Conclusion: The effort made in this study allowed us to approach the R-domains arrangement issue from a different point of view, sorting through the vast diversity of R proteins. Overall, many protein features were revealed and interesting new domain associations were found. In addition, insights on domain associations meaning and R domains modelling were provided.