#2492 Follow OBO/GO ID policy and do not change meaning of existing IDs

[Chris Mungall]

CHEBI from early 2014 had this:

[Term]
id: CHEBI:38234
name: DNA polymerase inhibitor
namespace: chebi_ontology
synonym: "DNA polymerase inhibitors" RELATED []
is_a: CHEBI:23924  ! enzyme inhibitor

which had no definition, yet the concept is clear and unambiguous from the label, and is easily recognized as corresponding to the inhibition of this activity:

id: GO:0034061
name: DNA polymerase activity

CHEBI has reused the same ID for a more specific concept:

[Term]
id: CHEBI:38234
name: EC 2.7.7.7 (DNA-directed DNA polymerase) inhibitor
def: "An EC 2.7.7.* (nucleotidyltransferase) inhibitor that interferes with the action of a DNA-directed DNA polymerase (EC 2.7.7.7)." []
...

If OBO/GO ID policy was followed, then CHEBI:38234 would have been obsoleted, and a new ID would have been introduced for the more specific class.

Incidentally, why is EC being used to define the hierarchy rather than GO? As CHEBI is an OBO Foundry ontology it should be setting a good example and reusing other OBO ontologies.

[Chris Mungall]

Apologies for the formatting. I think this gets it right:

before:

[Term]
id: CHEBI:38234
name: DNA polymerase inhibitor
synonym: "DNA polymerase inhibitors" RELATED [ChEBI]
is_a: CHEBI:23924  ! enzyme inhibitor

after:

[Term]
id: CHEBI:38234
name: EC 2.7.7.7 (DNA-directed DNA polymerase) inhibitor
def: "An EC 2.7.7.* (nucleotidyltransferase) inhibitor that interferes with the action of a DNA-directed DNA polymerase (EC 2.7.7.7)." []