PhytoPath: an integrated resource for comparative phytopathogen genomics

Food security has emerged as one of the most significant challenges for humankind in the 21st century. Food shortages, high energy costs, conflicting demands on crop production for biofuel generation and the soaring demand for food from east and south-east Asia are combining to drive food prices to their highest levels for many years. A significant constraint on crop productivity is disease, which accounts for 10-20% losses in yield every year. Controlling plant diseases furthermore represents a significant cost to farmers both in time and resources. The development of new durable disease control strategies that can be deployed at low cost therefore represents one of the best means of ensuring sustainable food production. Plant pathogens (and other species, including the plants they afflict) are increasingly studied through the use of high throughput, automated experimental approaches that generate large quantities of data. For over ten years, the determination of the sequence of complete genomes (that is, all the information that determines the heritable characteristics of a species) has been possible. More recently, advances in technology have reduced the costs of genome sequencing drastically, and made possible the determination of individual genome sequences (thus allowing the sampling of populations to determine their characteristics). Similar improvements in technology have increased the quantity of data produced describing the expression of genes and proteins in a variety of experimental conditions. However, while public repositories exist for certain types of these new experimental data, there is no integrative resource available that unifies these to facilitate their interpretation. In the absence of such a resource, there is (at worst) a danger that data generated by new technologies is lost; or alternatively that every scientist wishing to exploit such data has to tediously and wastefully integrate and correct information from different data sets. For many scientists, the determination of a coherent, up-to-date body of data from different experiments is a near-impossible challenge, and a distraction from the challenge of using such information to solve real scientific problems. The Ensembl software platform comprises a suite of tools for the analysis, integration and display of data from complete genomes. It includes modules for the handling of population-wide genome variation amongst individuals, and the evolutionary comparison between species. The platform has been used to capture genomes from many species including vertebrates and plants. We now propose creating PhytoPath, a new resource based on Ensembl technology to capture data from phytopathogen genomes, in response to the increased interest in food security and the concomitant increase in high throughput data available for pathogens of interest. PhytoPath will be run by the EBI, Europe's leading bioinformatics service centre, but will take its scientific direction from members of the UK phytopathogen research community, who are directly engaged in producing and exploiting these data. The use of the Ensembl platform is not only cost-effective (taking advantage of solutions already developed for use in other contexts), but also offers the exciting prospect of providing access to host, pathogen (and vector) genomes through a common interface. Particularly, this will facilitate the development of a new type of resource, correlating phenotype (i.e. the symptoms of pathogen-mediated disease) in with genotype (i.e. individual genome sequence) in both host and pathogen. The leading current resource for plant disease phentoypes is the pathogen-host interactions database (PHI-base), maintained by Rothamsted Research. We will develop a new interface for supervised community curation of PHI-base and integrate PHI-base tightly within PhytoPath to ensure that the pathogen phenotype can be studied in its genomic context.

We will establish a resource (PhytoPath) for high throughput molecular biology data from important phytopathogens and pathogenic phenotypes based on the Ensembl software infrastructure for genome analysis and display and PHI-base, the leading resource describing phenotypes of pathogenic infection. Information will be stored in a relational database and made available through a number of public interfaces, including a genome browser, a query optimised data warehouse, and bulk data download. Services will be operated as an integrated part of the EBI's suite of public services, and integrated with other services offering access to genome-scale data from other species (e.g. the plant hosts of pathogen-mediated disease). PhytoPath will be run by a management board comprising key members of the U.K. phytopathogen research community, and will initially prioritise the incorporation of data from selected fungal and oomycete pathogens of particular interest in the U.K. Data types of interest to PhytoPath include genome sequence, variation information, functional and regulatory assays, ESTs, transcriptomic and proteomic data. Priority species for inclusion in the first release are Magnaporthe grisea, Mycosphaerella graminicola, and Phytophthora infestans. Subsequently, data will be selected for incorporation according to the current research priorities of the community. We will develop methods for population-scale variation analysis, and comparative genomic analysis between pathogenic and related non-pathogenic species, building the domain specific expertise of project partners); and include the results in each release of the database. We will also develop a new interface to support annotation of host-pathogen interactions by community users, and develop a new interface linking genotypes in host and pathogen species with the disease phenotypes.

PhytoPath will be extensively used by the agricultural biotechnology/agrochemical industry, where there is considerable demand for a unified, centrally-curated database of genomic information for key plant pathogenic species. The development of an integrated resource aligning genomic, transcriptomic and proteomics data from both pathogen and host will provide a platform for researchers wishing to apply a systems approach to pathogen-mediated disease. In a wider sense, the potential benefits of the application of ultra-high throughput sequencing technologies to plant pathogens for target identification, include the potential for significant reduction in the cost of disease control and increases in achievable yields (benefiting both food and biofuel production). The economic and quality-of-life benefits of such advances are massive. While the development of PhytoPath is in itself only a small step towards a new green revolution, the proper management and integration of genome scale data is crucial for its correct interpretation and downstream exploitation.