Realizing the potential of reference ontologies for the semantic web

点击翻译按钮获取中文摘要

DESCRIPTION (provided by applicant): This proposal is motivated by the need to integrate vast amounts of complex biomedical information into a holistic understanding of biological function, and to apply that understanding to improved health care. The semantic web (SW), which is envisioned as a distributed network of knowledge layered on the web, is emerging as an essential part of the solution to this need since it will help to ensure that diverse sources of data mean the same thing. A critical component of the SW will be networks of interlinked ontologies, each of which is a collection of entities and relationships describing a domain of interest. At present most ontologies can be called application ontologies because they are designed for specific applications. However, these ontologies do not scale up, they often overlap with each other, and they are very difficult to link together because of incompatible knowledge models. Reference ontologies, such as our Foundational Model of Anatomy (FMA), are an emerging ontology type that have the potential to provide a conceptual foundation for linking together these diverse application ontologies into the SW. However, because their scope is so large reference ontologies are highly complex, and are very difficult for application developers to use. It is the purpose of this proposal to make such reference ontologies easy to use by developing methods for deriving application ontologies from them, so that they may fulfill their potential as a foundation for the semantic web. We will do this by adapting and extending research from the database community in order to 1) create application ontologies as views over reference ontologies, and 2) embed these views as query able web services. In developing these methods we will pursue the following specific aims, in collaboration with the National Center for Bio-ontology (CBio) and the UW computer science department: 1) investigate and develop view-based approaches for mapping between reference and application ontologies; 2) design and develop software tools that implement these approaches; and 3) develop graphical interfaces that allow end-users to specify these mappings. The tools we develop will be integrated as one component of the CBio Bio-Portal framework for accessing interlinked ontologies and data. Their development and evaluation will primarily be driven by the need to integrate data and computational models describing cardiac function.