From GO Wiki
Jump to navigation Jump to search


This page describes how logical definitions for "Regulation of Biological Attribute/Quality" terms in GO are created - for example, regulation of actin filament length



Terms under the GO regulation of biological quality node have simple logical definitions:

id: GO:0030832  ! regulation of actin filament length
intersection_of: GO:0065007  ! biological regulation
intersection_of: regulates OBA:0000011  ! actin filament length

The complexity is pushed to a separate ontology of attributes (currently in the GO extensions repo, but may be moved to a separate ontology maintained by multiple groups)

id: OBA:0000011
name: actin filament length
is_a: OBA:0000057 {is_inferred="true"} ! cellular component size
intersection_of: OBA:0000001  ! biological attribute
intersection_of: affects_quality PATO:0000122  ! length
intersection_of: attribute_of GO:0005884  ! actin filament

TermGenie templates

There is no live TG template for making new "regulation of <entity><quality>" terms yet.

However, there is a TG instance for the bio-attribute ontology:

Attribute Ontology

The attribute ontology is defined according to the general pattern

 'biological attribute' AND
    'affects quality' some PATO:nnnn AND
    'attribute_of' some <ENTITY> AND
    'occurs_in' some <OBJECT>

This is experimental, and may change

Note that the attributes are not themselves qualities. We use the occurs_in relation, which renders them occurrents, which is useful for describing a quality that changes over time. This also brings uniformity to the GO regulation pattern - the GO term "regulation of biological quality" is really regulating the progression of that quality through time.


Counts are represented like this:

id: OBA:0000112
name: myofibril number
is_a: OBA:0000001 ! biological attribute
intersection_of: OBA:0000001  ! biological attribute
intersection_of: affects_quality PATO:0000070  ! count
intersection_of: attribute_of GO:0030016  ! myofibril

This is a departure from the previous pattern whereby we avoided terms like "count" which are not true qualities in the BFO sense since they do not inhere_in anything - we avoid the use of the inheres_in relation, although this may be implicit in the connection between the affected quality and the target of the attribute_of relation.


The location of the attribute is represented as follows:

id: OBA:0000041
name: calcium ion concentration in cytosol
is_a: OBA:0000040 ! calcium ion concentration
intersection_of: OBA:0000001  ! biological attribute
intersection_of: affects_quality PATO:0000033 ! concentration of
intersection_of: attribute_of CHEBI:39124 ! calcium ion
intersection_of: occurs_in GO:0005829 ! cytosol

Curation of attribute ontology

Should largely be automated. New terms should be added via TG.

  • Classification should be automatic. Avoid asserting is_a
  • Do not edit any term in VT or TO namespace. These are merged automatically.

When new terms are added to VT or TO they may end up duplicating what is in OBA. In these cases, the OBA term will be obsoleted and replaced_by the relevant term in VT or TO.

We periodically bring in the latest VT and TO ontologies (plus associated logical definitions - http://code.google.com/p/phenotype-ontologies/ ). See the Makefile for details.