HemOnc.org is a large collaborative wiki providing evidence-based information on cancer regimens, along with their indications, components, and supporting literature. This project evaluates candidate data representation models for HemOnc.org content. Wiki content was obtained via a web-scraping algorithm. The resultant data was transformed into an OWL ontology, and it was subsequently converted to a custom relational database, a Neo4j graph database, and an OHDSI concept-relationship data model. We will compare and contrast these different formats.
Learning Objective: Understand the role of HemOnc.org in the oncology and informatics communities.
Understand the process of extracting information from a wiki into information models.
Be able to compare and contrast ontology models, graph databases, and relational databases for appropriateness with HemOnc.org data.
Zachary Moldwin (Presenter)
UIC College of Pharmacy
Harry Hochheiser, University Of Pittsburgh
Jeremy Warner, Vanderbilt University