Semantic Lookup Platform

When scientists publish research results, it is not unusual to use their own terminology and vocabulary. As science has become steadily more international and interdisciplinary in the last few decades, many different terms have been used that are semantically equivalent. The implementation of a semantic lookup platform encloses such semantic concepts that are used in the semantic layers to enable the downstream data integration in the IDSN project. The platform consists of two services - a terminology and a mapping service.

The terminology service (based on Ontology Lookup Service (OLS)) provides a web-based user interface for exploring and visualizing the vocabularies, and additionally, a flexible RESTful API to access these resources programmatically. It also provides a utility that allows to regularly update vocabularies with ease. Furthermore, the terminology service includes a search index for terms and synonyms with autocomplete functionality. To manage vocabularies, the terminology also uses a flexible configuration system. We modified the OLS software to support integration of terminologies and similar resources. The web frontend of current OLS software is optimized toward EBI branding, hence, it is not suitable for other institutions. Therefore, we created a default style so that institutions can easily integrate their branding. We also built a configurable Docker-based image to ease the deployment of new services.

The mapping service, based on Ontology Cross-reference Service (OxO), is the second service integrated into the semantic lookup platform whose key features are, finding mappings (cross references) between ontology classes, visualize mappings of different lengths and providing links to the associated terminologies inside the terminology service. Users can search a web-based interface, in which they can explore relationships between several ontologies and its terms. To visualize a brief summary of all mappings and ontologies contained in the service, the landing page displays an interactive graph with all ontologies that have cross references to others. A mapping between two ontologies exists if one ontology contains a term that is operational equivalent to a term in the second ontology or vice versa. Furthermore, the mapping service organizes its ontology data in a graph structure, which has the advantage that one can derive mappings of length greater than one. The service can display mappings up to the length of three. These mappings of different lengths help to give more insights on the term one considers. Additionally, the mapping service provides a RESTful API to retrieve data sources as well as mappings. Moreover, we have developed a Docker image one can modify, build, and use to deploy the service in a few easy steps.