Please indicate the label for the new term
ontology subset
Please provide a textual definition
a collection of classes that are grouped according to some purpose. analagous to a metaclass, but avoids the overhead of punning in owl
Please add an example of usage for that term
Please provide any additional information below. (e.g., proposed position in the IAO hierarchy)
subclassof: information artifact xref: obo-format:subsetdef
Comment #1
Posted on Feb 3, 2011 by Happy MonkeyThe definition isn't clear. How is it analogous to a metaclass? How does it avoid the overhead of punning in OWL. What exactly is the overhead of punning in OWL? If this is documented in OBO format somewhere, please include a pointer to that. Please provide an example of usage.
If I get the idea an instance of an ontology subset is related to classes by some annotation property? Are there requirements on the instances (like they should have documentation saying what the purpose is?
Comment #2
Posted on Feb 3, 2011 by Swift DogExample of usage: http://www.geneontology.org/GO.slims.shtml
The metaclass part is meant to be informative rather than normative, but if it doesn't help we can leave it out. There was discussion on this on the obo-format list probably makes sense to continue there. http://groups.google.com/group/obo-format/browse_thread/thread/17011df53049d3f7
Comment #3
Posted on Feb 12, 2011 by Happy Monkey(No comment was entered for this change.)
Comment #4
Posted on Mar 10, 2011 by Quick LionBackground
It is sometimes useful, for all sorts of purposes, to be able to define arbitrary subsets of the set of all terms in an ontology. Here are three examples
1. Defining a slim version of an ontology (GO has several).
2. Temporarily filtering out embargoed terms, defined based on unpublished work, from the public version of an ontology. This is something I've had to do a couple of times.
3. Defining subsets of terms that are allowed for curation. Databases often have internal rules for this that cut across the class hierarchy in ways that can't be supported by reasoner queries. Subsets can provide a useful way to reflect these rules so that appropriate terms can be extracted from an ontology for use in tool or pre-load checking of curation.
There are many other possible uses. It's the arbitrariness that makes this so useful.
Comment #5
Posted on Mar 10, 2011 by Quick LionComment deleted
Status: New
Labels:
Type-Term
Priority-Medium