A screenshot of the DiscoverySpace Explorer. The Explorer comprises one 'view', which describes the cross-section of data required by the user, and one or more datasets, which provide the initial starting content. The view is displayed in the left panel of the Explorer and graphically represents all navigable paths as a tree of sets cascading from the root class (much like the data model from Figure 2). Some properties have a literal object, such as a name, a sequence or a comment field. Others are links to associated resources such as genes, proteins or pathways. The view determines the content being displayed in the main table of the Explorer; each property in the tree has a checkbox that is used to include or exclude the property as a column. The right panel of the Explorer holds the main display. Each dataset added to the Explorer is represented by a tab containing a table. Each table displays the member resources of the selected dataset (and their weight values) as rows. The properties of the resources, and the properties of associated resources, are represented as columns, as determined by the view. The table, with its novel nested header, reflects the structure and the color-coding of the view in the left panel. If no view is specified by the user then a default view is created from the class of the dataset. All datasets in an Explorer session must have the same data type, and that class must be shared by the root node of the view. In the example above the user has constructed a view consisting of a path from the root set of tag sequences 'Up & down regulated pancreas' through virtual tags from Refseq to their counterpart Refseq genes. Additionally, the user has restricted the set of linked virtual tags to only those from human Refseq (multi-species joins are supported by the data model) and only those tags at position one (closest to the 3' end) on the sense strand. The result of the mapping operation, a set of human Refseq genes, can be selected and dragged out of the Explorer for further interpretation. The frequency, or weight, is displayed on the left hand side of each tag sequence.