ONTOFORCE hosts a public website www.disqover.com containing linked data from 140 di erent data sources in the bio-medical eld, often semi-manually curated en linked. The total volume of this data exceeds 7 billion triples. The software suite behind it, DISQOVER, is also available as a standalone application developed by ONTOFORCE to link and integrate data from di erent data sources. Customers can host it on-site for data from di erent private, often sensitive internal sources. A large portion of these customers are major pharmaceutical companies from around the world but ONTOFORCE is also active in other verticals.
In both cases the semantic web platform DISQOVER o ers linked data via a user-friendly interface. When a user starts a text query and selects a eld of interest from the results (for example a set of genes or diseases, in general a set of entities), a dashboard is shown speci c to that eld of interest with a number faceted search widgets. This allows the user to see a breakdown of the results in di erent dimensions, represented by di erent facets (for example in the eld of interest genes: breakdown by chromosome, by gene ontology function or by organism). The user can then drill down and lter using any of these facets. Also available are a limited list of the most relevant individual search results and an overview of all entities linked to the current result set. At each point in the search the user can follow the presented links to new dashboard to continue his or her search. At all times provenance of every piece of information is available. This feature set makes DISQOVER especially t to query a data knowledge graph in an enterprise setting.
To make the data in the single public data endpoint hosted by ONTOFORCE available at low e ort for our customers, ONTOFORCE developed a federated search in DISQOVER with a limited but very practical scope. The goal was to have a seamless integration of public data o ered by ONTOFORCE and private data in the user interface and to perform all searches within the scope of an HTTP request. This constraint excluded a solution built on SPARQL, so it was developed entirely in-house. The scope was particularly limited because we only have one federation endpoint (with ONTOFORCE public data) and we focused on the queries which are supported by our API and used by the GUI of DISQOVER. It allows the customer to link private information to information in the public endpoint and allows them to extend public information with private