Semantics of phenotype annotations

From phenoscape
Revision as of 01:02, 29 August 2011 by Jim Balhoff (talk | contribs)

In Phenoscape data annotation, we typically model a phenotype as a class expression describing a set of organisms, for example 'has_part some (caudal_fin and bearer_of some bifurcated)'. How precisely this phenotype class is related to a taxon has important consequences for the resulting inferences and thus queries for phenotypes across the taxonomy. This is most important when making annotations to higher-level taxa. How should these annotations propagate to sub-taxa?

This page assumes that taxa are represented as OWL individuals, and the taxonomy is a tree of subclade_of/contains_clade relationships. For example:

Ictalurus_punctatus subclade_of Ictalurus subclade_of Ictaluridae subclade_of Siluriformes

These are transitive, so that Ictalurus is a subclade_of Siluriformes. Individual organisms are related to taxa via member_of/has_member. A member_of a taxon X which is a subclade_of taxon Y is also a member_of taxon Y.

Types of phenotype annotations

Class generalization

(member_of value Siluriformes) subClassOf (has_part some (caudal_fin and bearer_of some bifurcated))

Ancestral state annotation

Observation annotation