Hyperontology for the biomedical ontologist
ā¢Download as PPTX, PDFā¢
1 likeā¢426 views
Presented at the 2011 ICBO Workshop on working with multiple biomedical ontologies. We present a framework for designing and interrelating ontology modules which are indvidually represented in different underlying logical formalisms.
Recommended
More Related Content
Viewers also liked
Viewers also liked (19)
Similar to Hyperontology for the biomedical ontologist
Similar to Hyperontology for the biomedical ontologist (20)
More from Janna Hastings
More from Janna Hastings (19)
Recently uploaded
Recently uploaded (20)
Hyperontology for the biomedical ontologist
- 1. ICBO, Buffalo, July 2011 Hyperontology for the Biomedical OntologistA Sketch and Some Examples Oliver Kutz1 Till Mossakowski1, 2 Janna Hastings3,4 Alexander Garcia Castro 5 Aleksandra Sojic6 1 Research Center on Spatial Cognition, University of Bremen, Germany 2 DFKI GmbH Bremen and University of Bremen, Germany 3 Chemoinformatics and Metabolism, European Bioinformatics Institute, UK 4 Swiss Centre for Affective Sciences, University of Geneva, Switzerland 5 University of Arkansas for Medical Sciences, USA 6 European School of Molecular Medicine, Milan; and University of Milan, Italy
- 2. Heterogeneous Formalisms Tuesday, July 26, 2011 2 Non-monotonic logic Fuzzy logic OBO HOL Rules Frames Modal logic Temporal logic OWL CL RDF OWL-EL OWL-RL Hyperontology for Bio-ontologists (WoMBO @ ICBO 2011)
- 3. Tuesday, July 26, 2011 3 Hyperontology for Bio-ontologists (WoMBO @ ICBO 2011)
- 4. Heterogeneous Content Tuesday, July 26, 2011 4 age etiology family history diet cell type Patient information Disease information genotype pathways Granularity social status symptoms Time Qualitative /Quantative Physiological information metabolic profile blood pressure Pathological / Canonical Hyperontology for Bio-ontologists (WoMBO @ ICBO 2011)
- 5. Modularity as desiderata Small modules: human understandable, re-usable, one level at a time, etc. Proving-in-the-small Large programs constructed from modules: Composition, Linkage, etc. Proving-in-the-large ļ Idea of a Module Interconnection Language Tuesday, July 26, 2011 5 Hyperontology for Bio-ontologists (WoMBO @ ICBO 2011)
- 6. Modularity at design-time Subject-specific modules Logic-specific modules (extensions) āMinimalā expressivity for a given topic Imports and complex interrelationships Tuesday, July 26, 2011 6 Hyperontology for Bio-ontologists (WoMBO @ ICBO 2011)
- 7. Hyperontologyis a framework for interrelating heterogeneous ontologymodules Tuesday, July 26, 2011 7 Logical translation Modular connections Integrated tools and reasoners Hyperontology for Bio-ontologists (WoMBO @ ICBO 2011)
- 8. Systematically linking ontology modules defined in different formalisms requires: A logic graph Fixed logic translations Tuesday, July 26, 2011 8 Hyperontology for Bio-ontologists (WoMBO @ ICBO 2011)
- 9. Common Algebraic Specification Language Standardised many-sorted first-order specification language Various extensions and sublanguages, including higher-order dialects, modal logic, OWL-DL; Supports structured specifications including: imports, hiding, renaming, union, extensions. Tuesday, July 26, 2011 9 Hyperontology for Bio-ontologists (WoMBO @ ICBO 2011)
- 10. Semantics of structured specifications Tuesday, July 26, 2011 10 Hyperontology for Bio-ontologists (WoMBO @ ICBO 2011)
- 11. HETCASL Extension of CASL for seamless combination of different logics Provides syntactic āsugarā Tuesday, July 26, 2011 11 Hyperontology for Bio-ontologists (WoMBO @ ICBO 2011)
- 12. HETS ā The Heterogeneous Tool Set Structured representations Reuse/independent development of modules Library of logics/formalisms supported, incl. OWL-DL Various provers connected: incl. OWLDL, first-order, higher-order, model checker Tuesday, July 26, 2011 12 Hyperontology for Bio-ontologists (WoMBO @ ICBO 2011)
- 13. Tuesday, July 26, 2011 13
- 14. Applications A sketch of some scenarios from bio-ontologies Tuesday, July 26, 2011 14 Hyperontology for Bio-ontologists (WoMBO @ ICBO 2011)
- 15. HOL as constraint Both the Sequence Ontology (SO) and the RNAO provide axiomatizations in first-order and higher-order logic to further constrain the semantics of the relationships that they use in OWL (not explicitly linked to the OWL version) Tuesday, July 26, 2011 15 Hyperontology for Bio-ontologists (WoMBO @ ICBO 2011)
- 16. Tuesday, July 26, 2011 16 Hyperontology for Bio-ontologists (WoMBO @ ICBO 2011) SIMULATION ANATOMY
- 17. Biochemical Structures Tuesday, July 26, 2011 17 Hyperontology for Bio-ontologists (WoMBO @ ICBO 2011)
- 18. Fullerenes Tuesday, July 26, 2011 18 Hyperontology for Bio-ontologists (WoMBO @ ICBO 2011) Cubic, 3-connected, planar graphs in which all faces are either pentagons or hexagons (5 or 6 atoms). Cubic 3-connected Planar (Kuratowski) Can be defined with Monadic Second Order Logic (MSOL)
- 19. Interrelating Ontologies Tuesday, July 26, 2011 19 Hyperontology for Bio-ontologists (WoMBO @ ICBO 2011)
- 20. Tuesday, July 26, 2011 20 Hyperontology for Bio-ontologists (WoMBO @ ICBO 2011)
- 21. Conclusions Biomedical ontologies are highly complex and need diverse formalisms for proper treatment The Hyperontology framework provides the āplumbingā to seamlessly integrate such formalisms Tuesday, July 26, 2011 21 Hyperontology for Bio-ontologists (WoMBO @ ICBO 2011)
- 22. Acknowledgements Funding DFG-funded collaborative research centre SFB/TR 8 `Spatial Cognition' The German Federal Ministry of Education and Research (Project 01 IW 07002 FormalSafe). Tuesday, July 26, 2011 22
Editor's Notes
- Although there is movement towards a standardisation on OWL as the central currency of ontology representation and exchange, diversity remains, since different profiles of OWL are good for different representation and reasoning tasks (and there is the usual tradeoff between complexity and scalability). OWLDifferent profiles of OWLDifferent logical languagesDifferent subjects require different levels of expressivity to adequately treat the subject matter The emphasis here is on different tools and syntaxes; differences in the underlying logic are treated in the next slide.
- The example of disease, different qualitative and quantitative data, and so on. Temporal data, etc. Granularity, epistemic differences. In particular, ontologies for complex diseases such as cancer have to deal with spatio-temporal het-erogeneity, combinations of qualitative and quantitative data, and missing links betweenphysiological and pathological data
- The benefits of modular design are well known in other engineering disciplines, including software engineering. To make modularity work for ontologies, however, requires the possibility of combining modules of different underlying formalisms.
- Features of hyper ontology framework: onto-logical translation graph
- Features of hyper ontology framework: different logical formalisms
- Features of hyper ontology framework: modularity
- Features of hyperontology framework: Implementation of HETS and HETCASL.
- DOLCE as an upper level ontology, and a treatment of mereologyUpper level ontologies in general can aim for a sophisticated axiomatization. Also, both the Sequence Ontology and the RNAO provide axiomatizations in first-order and higher-order logic to further constrain the semantics of the relationships that they use in OWL. But these axiomatizations are not explicitly linked to the OWL version.
- DICOM (Digital Imaging and Communications in Medicine) is a standard for handling, storing, printing, and transmitting information in medical imaging. It includes a file format definition and a network communications protocol. Biomedical example one: COMPUTATIONAL SIMULATION OF MECHANICAL PROPERTIES OF BONES When assessing the mechanical properties of bones, researchers use computational sim-ulations to evaluate stress and strain maps under several boundary and load conditions.Such evaluations involve clinical data, e.g. pathological conditions of the patients, andmechanical properties of the materials to be used. For describing a medical image it is often necessary to use several ontologies. Forinstance, Fig. 2(b) (left) illustrates the model for knee joints; hard tissue, e.g. Femurand Tibia bones, and soft tissue, e.g. Tibia and Femur cartilage, need to be identi-ed. The osteoarthritis of the patient requires shaving the cartilage injury as presentedin Fig. 2(b) (right). Such self-descriptiveness makes it possible for users to express complexqueries such as: `Retrieve knee joints images with cartilage injury.āFacilitating the execution of such a query requires the orchestration of several ontolo-gies, namely: Radiology, Anatomy, Pathology, CAD.
- Also the RNAO suffers from the problem of needing to represent cyclical structures in protein motifs
- ChEBIāGO, OBIThe existing OWL:import mechanism requires the full content ofboth ontologies be loaded into an application (such as Protege) in order to work with thecross-ontology denitions. The hyperontology framework allows us to bypass this problemwith its built-in support for modularisation, even across ontology languages.The hyperontology framework allows domain specicationssuch as the LO to eortlessly re-use parts of core ontologies such as ChEBI and even re-name or redene certain of their entities where needed. Also, more complex relationshipsbetween the ontologies' terms can be formalised in a heterogeneous ontology in the styleof Bridge Rules as they are known from distributed DL or E-Connections (see [12]).These classesare usually brought together via either OWL imports of the full ontologies (leading to asize explosion and the accompanying decrease in performance for reasoning) or simplyby \\slicing" the ontologies and putting together the classes on a need-to-have basis ac-cording to the MIREOT methodology [2]. This mechanism is facilitated by tools such asOntoFox [22], which allows users to input terms, fetch selected properties, annotations,and certain classes of related terms from source ontologies and save the results using theRDF/XML serialization of OWL. These hand-selected modules of external ontologies arethen brought manually into the target ontology through imports, and the procedure hasto be repeated every time the source ontology changes.
- Systems biology ontology (SBO): Combining mathematical expressions with ontologies. The types of questions that we might want to have answered relate to the ontological implications of the mathematical expressions rather than just to their answers, although some interesting work in extending concrete domains in this direction is surely taking place. SBO contains manydierent types of entities: material entities such as proteins and small molecules; processparticipation roles such as inhibitor and stimulant; mathematical laws such as rate lawsfor biochemical reactions; and types of mathematical model experiments such as discreteand continuous, and many more besides. The hyperontology framework would allow areformulation of the SBO as composed of modular units sourced from separate domainontologies, a highly desirable goal. Furthermore, of particular interest in the SBO is thatit captures complex mathematical relationships that can exist between biological entitiesin dynamic conditions. In SBO, these relationships are currently expressed in MathML.It is an open challenge to expose some of the relational information encoded in the SBOmathematical equations to ontological reasoning.