The foundational model of anatomy in OWL: Experience and perspectives
Introduction
Life sciences have a long tradition of controlled vocabularies. Extensive terminologies, classifications and ontologies have been developed for many years in various biomedical domains. These resources have the potential to contribute to the Semantic Web for Life Sciences, but need to be adapted for it. A large library of biomedical ontologies has been developed in frames, often with Protégé [12]. As OWL is the W3C recommended standard for ontologies [1], converting frame-based ontologies to OWL becomes an important need. Representing ontologies in OWL provides several advantages. Once converted to OWL, ontologies currently developed with frames become easier to integrate with other ontologies and can be used as resources for the Semantic Web. Interoperability of Web ontologies is important for shared use across different biological and medical domains, as expected for example from the Open Biomedical Ontologies (OBO) library. Also of interest is OWL higher expressiveness, and precise formal semantics. Another important advantage of OWL is the existence of powerful reasoning services, based on its underlying description logics. Several major ontological and terminological resources in biomedicine have been recently converted to OWL DL, including the Medical Subject Headings (MeSH) [8], the Gene Ontology™ [9] and the National Cancer Institute Thesaurus [10]. The conversion of other ontologies to OWL has also been investigated, e.g., the UMLS® Metathesaurus® [2] and Semantic Network [11]. Our long term goal is to provide a Web service assisting the conversion of frame-based and OBO ontologies to OWL. Meanwhile, the present study investigates the conversion of a large frame-based ontology into OWL and the reasoning services enabled by this conversion.
The frame-based ontology under investigation is the Digital Anatomist Foundational Model of Anatomy (FMA). It was converted from Protégé 2.1 to OWL DL. The FMA is the most comprehensive ontology of human ‘canonical’ anatomy [3]. The version used in this study, dated of July 2004, contains 70,169 concepts and more than 1.5 million relations. The FMA was selected for two major reasons. First, anatomy plays a prominent role in biomedicine and many biomedical ontologies and applications refer to anatomical ontologies. As its authors claim, the FMA is “a reference ontology in biomedical informatics for correlating different views of anatomy, aligning existing and emerging ontologies in bioinformatics …” [3]. Anatomy, together with Gene and Disease reference ontologies, constitute the backbone of the future Semantic Web for Life Sciences. Second, representing the FMA into OWL poses a real challenge from a knowledge representation perspective. It is important to investigate if OWL DL, which is a first order language, has sufficient expressiveness to represent what was originally represented with frames and metaclasses in Protégé. The capacity of OWL editors (e.g., Protégé OWL) and reasoners (e.g., Racer) to deal with the sheer size and complexity of the FMA and the scalability of OWL DL inference techniques to such a large biomedical ontology must be evaluated. Rather than OWL Full used in [5], we selected OWL DL because a main component of our study is to investigate the benefits of a DL representation over frames in terms of reasoning supported by the underlying description logic. OWL DL provides completeness and decidability of the interesting reasoning problems (satisfiability and subsumption) and supports consistency checking and automatic classification. OWL DL reasoners are available (e.g., Racer [16] and Pellet [17]). In contrast, OWL Full is undecidable, offers no computational guarantees and lacks suitable reasoners.
The rest of this paper is organized as follows. The method used to automatically convert the FMA from Protégé 2.1 into OWL DL is first presented (Section 2). Our experience in reasoning with OWL is reported next (Section 3). The choices of conversion, as well as possible perspectives for the FMA and open questions for large-scale ontologies of the future Semantic Web are finally discussed (Section 4).
Section snippets
Conversion to OWL DL
As DLs and frames share the same object paradigm, it might be thought that converting a Protégé frame-based1 ontology into OWL is straightforward and could be achieved by a simple export function mapping Protégé primitives to OWL
Reasoning with OWL
Reasoning with OWL proved to be a real challenge, due to the sheer size and complexity of the FMA. Processing the full-fledged FMA in OWL DL raised significant inference problems in terms of time and memory requirements. For this reason, an incremental approach to investigating reasoning services was adopted.
Discussion and perspectives
This study is a first step towards the representation of the FMA in OWL. Several issues remain open and different perspectives shall be considered in the future.
Acknowledgments
This research was supported in part by the Intramural Research Program of the National Institutes of Health (NIH), National Library of Medicine (NLM), while Pr. Christine Golbreich and Dr. Songmao Zhang were visiting scholars at the Lister Hill National Center for Biomedical Communications, NLM, NIH. Our thanks to Volker Haarslev and Ralf Möller for their valuable advice and help on OWL and Racer, to Cornelius Rosse, José Mejino and Todd Detwiler for the FMA, and to other helpful advice.
References (17)
- et al.
A reference ontology for biomedical informatics: the Foundational Model of Anatomy
J. Biomed. Inform.
(2003) - et al.
The National Cancer Institute's thesaurus and ontology
J. Web Semantics
(2003) - M. Dean, G. Schreiber, S. Bechhofer, F. van Harmelen, J. Hendler, I. Horrocks, D.L. McGuinness, P.F. Patel-Schneider,...
- et al.
Usability of expressive description logics—a case study in UMLS
Proc. AMIA Symp.
(2002) - et al.
Pushing the envelope: challenges in a frame-based representation of human anatomy
Data Knowledge Eng. J.
(2002) - et al.
Challenges in converting frame-based ontology into OWL: the Foundational Model of Anatomy Case-Study
Proc. AMIA Annu. Symp.
(2005) - et al.
Migrating the FMA from Protégé to OWL
- et al.
What reasoning support for Ontology and Rules? The brain anatomy case study
Cited by (96)
Ontology Based Data Access in Statoil
2017, Journal of Web SemanticsFrom frames to OWL2: Converting the Foundational Model of Anatomy
2016, Artificial Intelligence in MedicineCitation Excerpt :There has therefore been an increasing demand, both by us and by other FMA users, to convert the FMA to OWL so that it may more easily contribute to the global knowledge base. The conversion of the FMA into OWL has been attempted several times in the past [7–11]. Our original intension was to extend the implementations from one or more of these earlier efforts.
Towards the taxonomic categorization and recognition of nanoparticle shapes
2015, Nanomedicine: Nanotechnology, Biology, and MedicineFormalizing MedDRA to support semantic reasoning on adverse drug reaction terms
2014, Journal of Biomedical InformaticsConsequence-based and fixed-parameter tractable reasoning in description logics
2014, Artificial IntelligenceQuery rewriting under query refinements
2014, Knowledge-Based SystemsCitation Excerpt :Then, access to the data is performed via conjunctive queries (CQs) and the computed answers reflect both the stored data as well as the knowledge in the ontology providing so-called Ontology Based Data Access (OBDA) [34]. Ontologies also play an important role in a number of different scenarios like data integration [7,22], biomedical applications [15,12], and more. An important family of formalisms for constructing ontologies, primary due to the fact that they constitute the logical underpinnings of the Web Ontology Languages OWL [19] and OWL 2 [11] are Description Logics (DLs) [3].