Difference between revisions of "Individual-based taxonomy"

From phenoscape
 
(8 intermediate revisions by the same user not shown)
Line 1: Line 1:
Representing evolutionary lineages using ontological individuals has some modeling advantages, and may be more realistic, compared with using a simple class hierarchy. We described some of the rationale in [http://ceur-ws.org/Vol-833/paper82.pdf these ICBO conference proceedings]. For the purposes of the Phenoscape KB,  this model supports specification of the [[Semantics_of_phenotype_annotations|semantics of particular phenotype annotations]]: the phenotype applies to all members of a clade, known for some members, or known for the ancestor of the members.
+
Representing evolutionary lineages using ontological individuals has some modeling advantages, and may be more realistic, compared with using a simple class hierarchy. We described some of the rationale in [http://ceur-ws.org/Vol-833/paper82.pdf these ICBO conference proceedings]. For the purposes of the Phenoscape KB,  this model supports specification of the [[Semantics_of_phenotype_annotations|semantics of particular phenotype annotations]]: the phenotype applies to all members of a clade, is observed for some members, or is applied to the ancestor of the members.
 +
 
 +
In the KB build process, the VTO taxon class hierarchy is translated into a relationship graph. For every class relationships such as
 +
 
 +
Class: Ictaluridae
 +
    SubClassOf: Siluriformes
 +
 
 +
individuals with the same IDs (using OWL punning) are created with a relationship such as:
 +
 
 +
Individual: Ictaluridae
 +
    Facts: subCladeOf Siluriformes
 +
 
 +
The properties used in the evolutionary lineage model are currently held in the [http://purl.org/phenoscape/vocab.owl Phenoscape KB vocabulary ontology] (a holding place for KB-specific and preliminary terms). These include the inverse property pairs ''subCladeOf/containsClade'' (taxon to taxon) and ''memberOf/hasMember'' (organism to taxon/taxon to organism). ''subCladeOf'' is transitive, and also a member of a given taxon is inferred to be a member of all taxa containing that taxon (using a property chain).
 +
 
 +
==Comparison with CDAO tree model==
 +
The [http://www.evolutionaryontology.org/cdao Comparative Data Analysis Ontology] (CDAO) includes classes and properties for modeling phylogenetic trees which are different from the ones used here. We distinguish between models of information artifacts (e.g. a phylogeny produced by a particular analysis), such as is described by CDAO, and models of evolutionary lineages in nature. Obviously they are tightly related; the former is used to make hypotheses about the latter. However, the type of reasoning we would like to enable for each may be slightly different. The usefulness of this distinction is not yet well developed, however one example is that for a phylogenetic tree information artifact we can state that we know all of the nodes descending from a given node; i.e. we can close some of the open world.
  
 
[[Category:Reasoning]]
 
[[Category:Reasoning]]
 
[[Category:Ontology]]
 
[[Category:Ontology]]

Latest revision as of 20:31, 8 July 2013

Representing evolutionary lineages using ontological individuals has some modeling advantages, and may be more realistic, compared with using a simple class hierarchy. We described some of the rationale in these ICBO conference proceedings. For the purposes of the Phenoscape KB, this model supports specification of the semantics of particular phenotype annotations: the phenotype applies to all members of a clade, is observed for some members, or is applied to the ancestor of the members.

In the KB build process, the VTO taxon class hierarchy is translated into a relationship graph. For every class relationships such as

Class: Ictaluridae
    SubClassOf: Siluriformes

individuals with the same IDs (using OWL punning) are created with a relationship such as:

Individual: Ictaluridae
    Facts: subCladeOf Siluriformes

The properties used in the evolutionary lineage model are currently held in the Phenoscape KB vocabulary ontology (a holding place for KB-specific and preliminary terms). These include the inverse property pairs subCladeOf/containsClade (taxon to taxon) and memberOf/hasMember (organism to taxon/taxon to organism). subCladeOf is transitive, and also a member of a given taxon is inferred to be a member of all taxa containing that taxon (using a property chain).

Comparison with CDAO tree model

The Comparative Data Analysis Ontology (CDAO) includes classes and properties for modeling phylogenetic trees which are different from the ones used here. We distinguish between models of information artifacts (e.g. a phylogeny produced by a particular analysis), such as is described by CDAO, and models of evolutionary lineages in nature. Obviously they are tightly related; the former is used to make hypotheses about the latter. However, the type of reasoning we would like to enable for each may be slightly different. The usefulness of this distinction is not yet well developed, however one example is that for a phylogenetic tree information artifact we can state that we know all of the nodes descending from a given node; i.e. we can close some of the open world.