7. KnetMiner Support for new data (concept) types in KnetMaps

When integrating novel data resulting in the presence of new conceptClasses and relationTypes in the knowledge network oxl, the following customizations need to be undertaken to render the new concepts in the KnetMiner user interface, i.e., in KnetMaps (Network View), Evidence View legend and the dynamic, interactive legends for Gene View and KnetMaps:

• For KnetMaps, all visual attributes (color, shape, size) for concept and relations are defined in the cyjsJSON Exporter plugin module in the ondex-knet-builder platform for data integration. This module generates JSON, after a KnetMiner search, when a gene is selected to view its sub-network in KnetMaps (Network View). So, for any new concept classes, e.g, DGES, PO and COEXPS, users must add them to ConceptType.java within the cyjsJSON module and add any new concept shape and color options there similarly as well. Permitted shapes for later visualization in KnetMiner can be seen in ConceptShape.java. Note: These .java classes are Enums (containing constants) which are used as config files, without the overhead of parsing separate XML config files and with added type-safety. Once the concept and relation types and their visual attributes are added here, they are exported as JSON whenever KnetMaps is invoked within KnetMiner.

  Note: If these parameters are not defined, the exporter plugin will use default values to customize the new concepts (shape: roundrectangle, color: greenYellow, like for Phenotype concepts) and relations (color: black).

• Once the above is added and KnetMiner server .jar is deployed, users can visualize these new concept and relation types within KnetMiner. The next step is to add images of these new concept types in the image repository in KnetMiner client. This can be easily done via a KnetMiner search and then opening a sub-network for a gene in KnetMaps from where users can extract the concept's icon (e.g., as a png image) and add it to the above image folder. The Gene View and KnetMaps legend are both generated dynamically based on what concept types are present in the current search or the current sub-network being visualized but look at this image repository for images to add to the legend. Without adding these images, the legend will show concept type (and count) but not the image.