Fig. 3

The process used for building the prototype RDF search system to answer the Penn Medicine Biobank case/control competency question. 1. Semantic Modeling-Ontology models are developed to model the semantics of the relational data and any OBO ontologies that are relevant to the data sources and potential queries. 2. Data Mapping and Instantiation-The models developed in step 1 are used to write mapping files to concretely map the relational data as RDF. Software tools to use these maps to instantiate the relational data as RDF data. 3. Domain Knowledge Linking-The instantiated RDF data and any relevant OBO Foundry Ontologies are loaded into a graph database. 4. Querying and Testing-Queries over the graph data can be created by referencing the OBIB model. To test, equivalent queries against the graph data and relational data are constructed and run to ensure data correctness