Recent advances in digitization and dissemination of huge amounts of biomedical data make it more difficult for the user to successfully search and retrieve information [2]. Ontologies [3] provide a semantic layer, which facilitates data management and browsing. They conceptualize the domain of application by representing it as a set of concepts and relations between them. Depending on the expressivity of the formal ontology language used (RDF Resource Description Framework, RDF [4], Resource Description Framework, RDFS [5], Web Ontology Language, OWL [6]), the relations may range based on the restrictions applied to them, i.e., from hierarchical to more complex restrictions. Concepts of ontologies drive query formulation over the content stored in repositories or Knowledge Bases (KBs) and help in indexation by providing formalized classification of data and information. KBs and repositories that use ontologies as their semantic backbones of classification and organization of stored data and information can be explored by navigating ontologies that they rely on [7]. OWL language provides a rich set of constructors to model complex relationships between concepts. For example, reference biomedical ontologies heavily use existential restrictions to express parthood and functional relations of anatomical entities [8]. OWL 2 EL [9] profile (a subset of OWL 2 language) is particularly popular in the biomedical domain and captures most of the expressivity required for biomedical ontologies while keeping the polynomial computational complexity for DL reasoning tasks (e.g., ontology consistency, class expression subsumption, and instance checking). Consequently, the ontology visualization and interactive ontology exploration techniques should support the expressive power of an underlying ontology language to capture adequately the variability of the domain knowledge encoded. This deliverable describes Grontocrawler [1], a set of interactive ontology tools developed to support interactive semantic exploration of the Biomedical Knowledge Space pertained to the MSH domain. The Biomedical Knowledge Space is formalized and implemented in the MSH ontology (D8.2) following the modularized ontology design. Processing and visual exploration of ontologies with Grontocrawler are performed by transforming OWL 2 EL ontologies into directed graphs, in which nodes represent concepts or individuals and edges represent either explicitly encoded or inferred relations between concepts or individuals. Identification of edges is done by studying the concept dependence in the ontology expressed as a set of terminological (TBox) or assertional (ABox) Description Logic (DL) [10] axioms and implemented directly on RDF (OWL's data model) triples. The resultant graph representation of the ontology allows the re-use of the algorithms from the Graph Theory to accomplish the three tasks: network analysis, modularization and visual exploration (please refer to related work section for an overview of these tasks). The three tasks are then applied on the MSH Ontology to provide web-based exploration of the MSH knowledge space, as exemplified in the "Use-cases" section.
Multi-scale biological modalities for physiological human articulation
A Agibetov;I Banerjee;C E Catalano;M Spagnuolo
2015
Abstract
Recent advances in digitization and dissemination of huge amounts of biomedical data make it more difficult for the user to successfully search and retrieve information [2]. Ontologies [3] provide a semantic layer, which facilitates data management and browsing. They conceptualize the domain of application by representing it as a set of concepts and relations between them. Depending on the expressivity of the formal ontology language used (RDF Resource Description Framework, RDF [4], Resource Description Framework, RDFS [5], Web Ontology Language, OWL [6]), the relations may range based on the restrictions applied to them, i.e., from hierarchical to more complex restrictions. Concepts of ontologies drive query formulation over the content stored in repositories or Knowledge Bases (KBs) and help in indexation by providing formalized classification of data and information. KBs and repositories that use ontologies as their semantic backbones of classification and organization of stored data and information can be explored by navigating ontologies that they rely on [7]. OWL language provides a rich set of constructors to model complex relationships between concepts. For example, reference biomedical ontologies heavily use existential restrictions to express parthood and functional relations of anatomical entities [8]. OWL 2 EL [9] profile (a subset of OWL 2 language) is particularly popular in the biomedical domain and captures most of the expressivity required for biomedical ontologies while keeping the polynomial computational complexity for DL reasoning tasks (e.g., ontology consistency, class expression subsumption, and instance checking). Consequently, the ontology visualization and interactive ontology exploration techniques should support the expressive power of an underlying ontology language to capture adequately the variability of the domain knowledge encoded. This deliverable describes Grontocrawler [1], a set of interactive ontology tools developed to support interactive semantic exploration of the Biomedical Knowledge Space pertained to the MSH domain. The Biomedical Knowledge Space is formalized and implemented in the MSH ontology (D8.2) following the modularized ontology design. Processing and visual exploration of ontologies with Grontocrawler are performed by transforming OWL 2 EL ontologies into directed graphs, in which nodes represent concepts or individuals and edges represent either explicitly encoded or inferred relations between concepts or individuals. Identification of edges is done by studying the concept dependence in the ontology expressed as a set of terminological (TBox) or assertional (ABox) Description Logic (DL) [10] axioms and implemented directly on RDF (OWL's data model) triples. The resultant graph representation of the ontology allows the re-use of the algorithms from the Graph Theory to accomplish the three tasks: network analysis, modularization and visual exploration (please refer to related work section for an overview of these tasks). The three tasks are then applied on the MSH Ontology to provide web-based exploration of the MSH knowledge space, as exemplified in the "Use-cases" section.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
