Ontology types

To identify entities in a portion of text and type them, we use the knowledge sources (terminology, thesaurus, and ontology) introduced in Section 31.3.2. The three levels are used by the following procedures ... 

The ontology that underlies the information extraction and annotation process is solely based on taxonomic relationships. We intend to enrich our ontology with typed relationships. We are currently evaluating how typed relationships can extend the functionality of the UltraLink and how the expressivity for our ontology impacts the computational complexity of formal reasoning [9]. 

A property-type F is ontologically reducible to a more fundamental property-type G is the possibility of something s being F is constituted by a recombination of actual instances of G, but the possibility of something s being G is not constituted by a recombination of actual instances of F (12). 

The MAGE has built within it a controlled vocabulary that is used to standardize communication between data providers. However, MAGE can also be extended to encode other types of omic data beyond genomics, such as proteomic data, so long as a reference to the ontology or controlled vocabulary is provided. Although a description of the extension mechanism is beyond scope of this book, practitioners must become familiar with it in order to ensure their software products and applications will be able to accept any and all annotation data that may be submitted with the genomic data. 

It is worth mentioning that ontologies describe types of things, rather than things you would find in the universe, which are instances of the types. A building ontology would have entries for detached house, apartment, and shed, rather than 23 Acacia Avenue, Framley. For chemistry, a type would be water, and its instance would be a particular molecule of water in the sea or in a cup of coffee. 

The two foundational relationships between entries in an ontology are is a and part of. The is a relationship is a taxonomic one thus, a dog is a animal, a spaniel is a dog, and so forth. Likewise, the part of relationship specifies that a hinge is part of a door, a door is part of a room, and so on. However, there is a subtlety. Not all parts can be described with the part of relationship, because if you say that type A part of type B, that means that all instances of A are necessarily part of some instance of B. So hydrogen is not part of methane. There are very few genuine cases of a part of relationship that describes chemical compounds. It is better to use a reciprocal relationship, has part, in which all type A has part some type B, or all methane molecules have some hydrogen atoms as parts. 

Each SOAP Web service is described in terms of Web Service Description Language (Booth and Liu 2007), which is a specification in XML. This enables data passed over the Web, say a sequence of characters such as CMRSGGCTRRYAC, to have its type specified according to an ontology, thus telling the code that that sequence of characters is a consensus DNA sequence rather than an author name or a geographical location. A RESTful Web service works in terms of HTTP requests and thus has potentially a less constrained syntax. 

J. Bard, S. Y. Rhee, and M. Ashbumer. 2005. An ontology for cell types, Genome Biology 6 R21. 

ABC ATP CDX CNT ENT EMT GIT GO NCE PEPT1 PMT QSAR SAR SLC ATP-binding cassette Adenosine 5 -triphosphate Caudal-type homeobox transcription factor Concentrative nucleoside transporter Equilibrative nucleoside transporter Epithelial-mesenchymal transition Gastrointestinal tract Gene ontology New chemical entity Di/tri-peptide transporter 1 Pharmacogenetics of Membrane Transporters Project Quantitative structure-activity relationship Structure-activity relationship Solute carriers (SLCs)... 

---