Presenter: Jim Brinkley, University of Washington Structural Informatics Group
Abstract: "Ontology", which is the latest term for work that arose from efforts in philosophy and AI knowledge representation, is currently a hot topic in biomedical research because of the belief that ontologies have the potential to help integrate the increasingly unmanageable amount of biomedical data and knowledge. Many institutes of the National Institutes of Health (NIH) are funding efforts in this area, the most notable being the National Center for Biomedical Ontology (NCBO) at Stanford.
In this talk I will describe a collaborative NIH grant we have with NCBO that is based on our own Foundational Model of Anatomy (FMA), one of a growing number of large "reference" ontologies that are meant to cover most of the knowledge in basic science fields like anatomy or physiology. The purpose of the grant is to develop methods for deriving "application" ontologies from one or more of these large reference ontologies so that the reference ontologies become practical for use in real-world applications, and so that they fulfill their potential to help integrate a growing number of application ontologies into the larger life sciences semantic web. Our approach is based on the database notion of a "view", where for an ontology the view is defined by a query over the underlying reference ontology(s), in this case using extensions to the SparQL semantic web query language. I will describe extensions we have made to SparQL to facilitate the creation of views, the use of these extensions in creating simplified views of the FMA and the National Cancer Institute Metathesaurus that are accessible as ontology web services, and the potential uses of these views in data and computational model integration. This is joint work with Stanford, UW Biomedical Informatics and UW CSE (Dan Suciu and Linda Shapiro).