PRGdb 2.0: towards a community-based database model for the analysis of R-genes in plants.

The Plant Resistance Genes database (PRGdb; http://prgdb.org) is a comprehensive resource on resistance genes (R-genes), a major class of genes in plant genomes that convey disease resistance against pathogens. Initiated in 2009, the database has grown more than 6-fold to recently include annotation derived from recent plant genome sequencing projects. Release 2.0 currently hosts useful biological information on a set of 112 known and 104 310 putative R-genes present in 233 plant species and conferring resistance to 122 different pathogens. Moreover, the website has been completely redesigned with the implementation of Semantic MediaWiki technologies, which makes our repository freely accessed and easily edited by any scientists. To this purpose, we encourage plant biologist experts to join our annotation effort and share their knowledge on resistance-gene biology with the rest of the scientific community.

INTRODUCTION

Plants protect themselves from disease by activating a broad array of defense responses that ultimately inhibit growth and spread of invading pathogens. The principal immune mechanism against pathogens in plants is mediated by resistance (R) gene (1). However, the essential components of this defense system remain still unknown even if it has been deeply investigated. High-throughput genomic experiments and plant genome sequences available in public databases are offering unprecedented opportunities to identify novel R-genes, to explore their function and their diversification process, to discover new resistance capacity, and ultimately to elucidate their mechanisms of interaction between pathogens and their plant hosts.

With the intention to facilitate research on this agriculturally important gene family, we launched in 2009 the Plant Disease Resistance Gene database, a comprehensive repository of R-genes across hundreds of plant species (2). PRGdb version 1.0 comprised a total of 16 844 gene entries, of which 73 were known and manual-curated R-genes (e.g. the ‘reference’ data set), 6308 were putative R-genes retrieved from NCBI Genbank, and 10 463 were putative R-genes computationally predicted from NCBI UniGene transcriptomic data using the in-house developed bioinformatics pipeline DRAGO (Disease Resistance Analysis and Gene Orthology) (2). In the last few years, several plant genome-sequencing projects have progressed rapidly. For instance, tomato (3), potato (4) and melon (5) genomes were recently completed, and thus providing the opportunity to discover additional R-genes. Therefore, the huge amount of sequence data not yet analysed for this gene family required a substantial update of our database for addressing the community requests.

In this manuscript, we present an update of the plant resistance gene database, PRGdb. In the release 2.0, our database has been expanded more than 6-fold to include useful biological information about a total of 104 459 R-genes (of which ∼90 000 newly annotated) from 233 plant species. Of these entries, 112 are manual-curated R-genes described in the literature to confer resistance to 122 different pathogens. All remaining genes were predicted by our DRAGO bioinformatics pipeline from sequences available in public transcriptomic and genomic databases (NCBI Genbank, NCBI UniGene, Phytozome) (6), as well as from gene models annotated in recent plant genome sequence projects, such as potato (4), tomato (3) and melon (5). Last but not least, we completely revised our web interface that is now based on Semantic MediaWiki technologies, which makes our repository freely accessible and easily editable by the scientific community. With this new wiki design, PRGdb aims to provide a forum for scientists working in this research field. Therefore, we do invite plant biology experts to join our efforts and help us on making PRGdb a more collaborative and comprehensive resource through their valuable contributions.

DATA UPDATE

To support the scientific community working on plant disease resistance area, we updated the PRGdb knowledge-base. In this release 2.0, the number of genes of previous version has grown more than 6-fold, new sections have been created and the domain analysis and the sequence classification of the candidate R-genes have been improved. To date, PRGdb 2.0 stores 112 ‘reference’ R-genes (40 more than the previous version), and 122 pathogen and 233 plant species (45 more than the previous version). The analysis with the DRAGO bioinformatics pipeline on 2012 NCBI ‘nr’ and UniGene databases yielded a total of 34 558 putative R-genes (20 000 more than the previous version). In addition to these traditional annotated categories, we included a new dataset predicted by DRAGO from genomic sequences available at Phytozome v8 (6), thus incrementing our repository of 68 509 putative R-genes from 31 plant sequenced genomes. Each putative R-gene sequence has been annotated with DNA and amino acid sequences, information about species and predicted protein domains. Noteworthy large-scale data released from PRGdb are listed in Table 1. Finally, another unique and important set of R-genes has been annotated by our team during the analysis of plant genome species recently sequenced and not yet present in the Phytozome database, such as tomato, potato and melon. For these species, a complete analysis of their R-genes panorama has been accomplished and a dedicated section of the database has been built to store, visualize, and consult these unique results (see wiki ‘metaspecies’ section). Table 2 shows a comparison of the R-gene classes identified in these agriculturally important species. Moreover, in this update, we include new identified R-protein domains, thus reaching a total of nine well-known domain types: the MLO (PFAM PF03094) (7), the RPW8 (8), and the GNK2 domain extracted from the antifungal protein Ginkobilobin2-1 (Interpro IPR002902) (9), and the already present CC (Coiled Coil), TIR (Toll-Interleukin like Region), NBS (Nucleotide Binding site), LRR (Leucine-Rich Region), KIN (Kinase) and Other (all other domains which have been described as conferring resistance through different molecular mechanisms) (2). The aforementioned new domain types helped to better classify the new candidate R-genes (Table 2). All data mentioned have been organized in new database sections as well-established repository for plant scientists. In particular, single gene pages are now available to directly visualize gene features allowing comparison of R-gene complements across entire plant lineages.

Table 1.

List of main entries present in PRG database

Entries	No.
Plants	233
Pathogens	122
Domain types	9
Gene classes	16
Manually curated R-Genes	112
Putative R-Genes, collected from NCBI Protein	9639
Putative R-Genes, predicted from NCBI UniGene	24 918
Putative R-Genes, predicted from Pythozome	68 498
R-genes from plant genome sequencing projects	1143
Total number of plant genes	104 310
Avirulence genes	23
Diseases	120

Table 2.

Putative R-genes retrieved in each sequenced species divided for class

Species	CN	CNL	Mlo-like	N	NL	Other	RLK	RLK-GNK2	RLP	RPW8-NL	T	TN	TNL	Unknown	Total
Aquilegia coerulea	19	145	23	89	69	0	0	52	282	1	1	0	0	1010	1691
Arabidopsis lyrata	1	41	15	101	150	7	0	38	198	3	37	1	205	771	1568
Arabidopsis lyrata subsp. lyrata	3	36	23	41	40	0	0	15	142	1	7	0	25	629	962
Arabidopsis thaliana	65	748	40	125	230	46	53	90	580	19	180	12	683	3195	6066
Areca catechu	33	295	31	126	481	2	7	183	751	1	228	0	471	3217	5826
Brachypodium distachyon	4	89	6	15	57	1	0	22	70	0	1	0	0	240	505
Brassica rapa	31	71	21	65	40	10	2	72	306	3	26	0	119	1121	1887
Capsella rubella	18	66	16	47	25	12	1	63	242	5	43	0	56	785	1379
Chlamydomonas reinhardtii	3	0	0	25	0	0	0	0	0	0	0	0	0	165	193
Citrus clementina	18	340	24	101	176	1	15	28	449	1	15	0	84	1396	2648
Citrus sinensis	31	203	42	131	224	0	35	59	477	3	33	0	76	1916	3230
Cucumis melo	0	2	0	0	0	2	0	0	0	0	0	0	1	0	5
Cucumis sativus	12	159	31	60	152	0	12	19	247	0	1	0	0	1047	1740
Ectocarpus siliculosus	0	0	2	15	0	0	1	0	0	0	0	0	0	118	136
Eucalyptus grandis	24	239	29	104	314	2	7	170	693	1	111	0	300	2332	4326
Eutrema halophilum	15	50	15	51	24	7	0	40	246	2	27	0	72	739	1288
Glycine max	37	132	60	190	281	0	8	141	803	8	78	0	229	3068	5035
Gossypium arboreum	7	170	20	56	75	0	0	36	232	1	1	0	0	827	1425
Linum usitatissimum	14	24	25	82	44	0	3	29	346	1	36	0	114	1227	1945
Malus x domestica	32	317	42	99	663	11	20	71	518	4	115	0	299	2811	5002
Medicago truncatula	15	98	29	105	357	0	3	58	297	4	95	0	291	1664	3016
Oryza barthii	0	12	15	71	100	0	0	19	195	1	9	0	33	721	1176
Oryza sativa	14	91	17	91	533	1	1	62	439	1	2	0	0	1548	2800
Panicum virgatum	37	332	24	115	532	0	2	64	478	0	3	0	0	2293	3880
Pennisetum glaucum	17	21	27	42	28	1	2	24	239	1	4	0	19	844	1269
Phaseolus vulgaris	18	114	29	69	227	0	3	67	306	2	18	0	114	1255	2222
Physcomitrella patens	7	6	21	89	38	1	0	0	225	1	2	0	6	770	1166
Populus trichocarpa	14	188	35	166	809	4	6	56	540	7	63	0	322	1946	4156
Prunus persica	8	132	22	69	138	0	0	30	273	1	19	0	131	981	1804
Ricinus communis	15	103	16	66	161	2	1	36	240	0	1	0	2	1224	1867
Selaginella moellendorffii	0	1	20	79	4	0	6	7	127	0	0	0	0	769	1013
Setaria italica	22	194	16	73	204	3	0	45	314	0	2	0	0	1143	2016
Solanum lycopersicum	2	26	8	9	41	1	0	4	45	0	1	0	3	407	547
Solanum tuberosum	0	7	5	4	71	0	2	3	5	0	8	0	4	288	397
Sorghum bicolor	5	13	16	94	320	3	0	45	298	1	5	0	0	1218	2018
Vitis aestivalis	14	102	23	61	98	0	8	25	261	1	13	0	29	1299	1934
Vitis vinifera	17	164	33	96	300	0	3	32	355	1	10	0	42	1268	2321
Volvox carteri	1	0	4	15	0	0	0	0	0	0	0	0	0	87	107
Zea mays	24	106	59	125	104	6	44	63	515	4	2	0	0	2330	3382

NEW FEATURES

Several scientific databases allow a direct interaction of users for sharing experimental and theoretical information. Most of them have a wiki structure and can contain specific or more generic data. Examples are Gene Wiki (10,11), WikiPathways (12), SubtiWiki (13), EcoliWiki (14), Tetrahymena genome database Wiki (15), Reactome (16) or SNPedia (17). Like these, PRGdb interface has been completely rewritten in a wiki system because in this way, it can be easier to get a larger scientific community be involved in the accumulation of data. We have used MediaWiki, the open-source software behind Wikipedia, to build this second version of the PRGdb with new features, a friendly interface, and new tools. Mapping our existing database structure into the wiki has been possible thanks to a set of Semantic MediaWiki extensions. This allows users to access and list different wiki pages according to multiple and flexible criteria and, at the same time, it offers a linking point for third parties to existing data via RDF triples attached to every page (18). By using Semantic technologies, PRGdb becomes another node of the linked data web (http://linkeddata.org/) On the other hand, it retrieves and links to resources such as Bio2RDF (19) (e.g. for retrieving additional species images) and potentially to other semantic empowered sites.

Other international plant databases use different models to interact with users. Among them, TAIR (20), SOLGenomics Network (21), Cucurbit Genomics Database (http://www.icugi.org) and Plant Organelles Database (22) are based on specific submission forms to contact database developers or systems based on information provided by emails; SOLGenomics Network has a forum section allowing users to interact as well. In our case, we take advantage of default wiki features such as change tracking and other social-like components, such as user pages, which show users activity and can be filled by plant disease resistance researchers. Moreover, apart from plain wiki editing for documentation purposes, several specific contribution entry points are also provided, so that registered users can suggest corrections or additions, such as related bibliography or even new reference genes. A group of reviewers, experts in the plant disease resistance field, will have the necessary permission to approve these contributions and, from that moment, these will be easily available and also linkable for third parties. In this way, PRGdb becomes a community database for plant scientists who could in turn contribute to this public resource collaboratively.

According with our final intent, wiki pages have been designed in a user-friendly version, enriched with interactive images and pages that summarize species with known R-genes (Figure 1). In the wiki pages, the user can find internal links connecting different sections of the website and improving accessibility to sequence, resistance gene data and related published information. New sections about plant and pathogens have been added, allowing users to quickly retrieve information from their species of interest. Some interactive images have also been included to facilitate navigation in the web site and to summarize results coming from data collection. Moreover, wiki pages are also intended to provide technical details about the pipelines adopted to predict R-genes and information coming from international scientific literature built collaboratively. In particular, notable scientist from all around the world that showed interest in our initiative will be asked to be in PRGdb review board. Reviewers will responsible for reviewing scientific accuracy of information. Members of primary specialty will be asked to review content areas that coincide directly with their expertise. Information in this section not only can be directly added by users, but can also be directly linked to other scientific databases. A variety of links to other database relevant to the topic are included in the initial menu.

Figure 1.

An overview of the wiki interface of the PRGdb version 2.0. (A) Home page; (B) BLAST results; (C) Clickable image of species annotated with R-genes; (D) Examples of metaspecies page, and two examples of wiki search results; (E) Reference R-gene wiki page; (F) Contributed wiki page.

Moreover, we have developed a new interface for BLAST search (23,24), implemented directly in the wiki system, so that results links are directly connected to gene wiki pages. All the sequences stored in the wiki are also provided as downloadable files, which are updated periodically from approved user wiki contributors.

FUTURE PERSPECTIVES

Newly sequenced plant genomes will be monitored from public resources, such as NCBI, Phytozome or any other specialized resource. Releases are expected to take place twice per year by integrating new data from our reference database plus the accumulated user contributions up to that moment. At the technical level, we will continue the employment of pipelines and algorithms to automate the management and classification of the sequence data and to better organize the content of the database. Connecting the various resources with phenotypic, genomic and molecular information will become a crucial task for mining results.

FUNDING

Italian Ministry of Education, University and Research (GenHort Project). Funding for open access charge: Department of Soil, Plant, Environment and Animal Production Sciences, University of Naples ‘Federico II’, via Università 100, 80055, Portici, Italy. Spanish Ministry of Economy and Competitiveness funded infrastructure technician support. Ref.: PTA2010-4446-I.

Conflict of interest statement. None declared.