Figure 3

Taverna Workflow 2. Graphical representation of a workflow using ArkDB, PubMed and GenBank web services captured from Taverna Workbench v1.7.0.0. An ArkDB accession number for a Marker (which could itself be retrieved from another ArkDB web service) is handed in as a parameter to the ArkDB fetchMarkerByAccession service. The result returned from this web service is again a DataSet document conforming to the GenomicMappingData schema, which can be parsed by the Taverna XMLSplitter, to retrieve the DataSet Element and the ExternalReferences Element, which is then parsed by the Taverna 'XPath from Text' java widget to retrieve the SourceIDs for any PubMed and GenBank external references (xrefs) in this result document. (An XPath expression to retrieve all GenBank IDs from the document being '//gmd:ExternalReference/gmd:DataSource/@idref [.="GenBank"]/../../gmd:SourceID'). These IDs are used to parameterize requests to PubMed and GenBank web services to retrieve the source records for these xrefs, using Taverna NCBI java widgets ('Get PubMed XML by PMID' and 'Get Nucleotide GBSeq XML'). In this case the Workflow outputs the original request document which details all of the information held in ArkDB about this Marker (its Mappings, and any Relationships to other objects such as Markers and Sequences) and also documents for any referenced PubMed citations or GenBank sequences referred to in the result document. The Workflow document is available as additional file 6: demoWorkFlow2.xml, and the resulting data document is available as additional file 3: demo2result.xml.