UK CropNet and CORBA

¹ John Innes Centre, Norwich Research Park, Colney, Norwich, NR4 7UH, UK

² NASC, Division of Plant Science (UP), University of Nottingham, Nottingham, NG7 2RD, UK

³ Scottish Crop Research Institute, Invergowrie, Dundee, DD2 5DA, UK

⁴ Department of Plant Breeding, Cornell University, Ithaca, NY, USA

⁵ Institute of Grassland and Environmental Research, Plas Gogerddan, Aberystwyth, Ceredigion, SY23 3EB, UK

The BBSRC-funded UK Crop Plant Bioinformatics Network (UK CropNet) is using comparative mapping data to bind its single species databases to form a single cohesive entity. We have developed a strategy for creating a distributed computing system, to link the UK CropNet species databases through ComapDB, a database holding comparative mapping data. This will enable us to make complex, real time queries for comparative genome analysis.
We have made a decision to use CORBA (Common Object Request Broker Architecture) as the "glue" between our databases, displays and querying tools. At present we have two prototype systems for the front and back ends of this environment. Our initial prototype CORBA system was developed as a proof-of-concept, to highlight the problems that such a system will present. It uses a JavaBean and the InfoBus to extract information from the individual UK CropNet databases, to collate the results of these cross-species queries and to channel them to the UK CropNet graphical displays. This worked well but was hindered by the lack of a standard CORBA interface to ACEDB. Consequently, we had to create test versions of the UK CropNet databases, using the DBMS Microsoft Access.
Recently, we have developed a prototype CORBA interface to ACEDB (CITA) to allow connection to a variety of ACEDB databases independent of their underlying model schemata. A number of demonstrations can be viewed which show various clients in action, a tace style interface with full text-based access, GFace which has full graphical functionality (including clickable versions of the original ACEDB displays) and new Java displays for maps, tables and trees.
At present our data models for distributing comparative data are simple. However we are developing more complex models in collaboration with our sister Animal Network (Roslin Institute, http://www.ri.bbsrc.ac.uk/bioinformatics/) and in line with recommendations to be made by the OMG's Life Sciences Special Interest Group.