The National Cancer Institute (NCI) from the US has built an integrated biomedical platform called caBIG in order to improve the US cancer research community’s access to key bioinformatics data and tools. The caBIG programme was established in order to allow researchers to more efficiently discover, share, process and integrate disparate clinical and research data, with an ultimate goal of accelerating cancer research.
The backbone of the caBIG infrastructure is caGrid – a grid-computing service-oriented environment that integrates various data and analytical services and leverages the combined strengths and expertise of participating organizations in an open and federated environment. caBIG services are categorized as legacy, bronze, silver and gold based on their ability to interoperate with other caBIG services. Legacy does not guarantee any interoperability with external software components or data resources; silver services comply with standardized semantic advertising, discovery and use of the service; gold level is currently being defined by caBIG as an extension to the silver that will enable full syntactic and semantic interoperability of services. Most of the services currently deployed at caGrid are silver.
Being able to link caBIG services into a workflow is one of the key requirements for caBIG users, in order for them to automate the design and running of their virtual experiments. Taverna Workbench (henceforth Taverna), a workflow design, management and execution system developed by the myGrid team (now based in Manchester, UK), has emerged as the primary candidate that will meet this need. The myGrid team has joined efforts with the caGrid workflow team from the US (funded by the NCI) in order to integrate caBIG services into Taverna and enable caBIG users to orchestrate data and analytical services in Taverna workflows. This also included integration with the caGrid’s GAARDS security framework for interaction with secure caBIG services.
As a side effect of the integration, caBIG services can be combined in mix-and-match style workflows with other existing third party services already accessible from Taverna.
This paper describes the details of Taverna-caBIG integration – the extensions added to Taverna in order to be able to discover and invoke caBIG services from Taverna workflows.