Guidelines for cross-reference types

[pgaudet]

Splitting this discussion from
#28558 (comment)

We want to have guidelines for cross-reference types, in particular narrow versus related.
We have some guidelines here: https://wiki.geneontology.org/Guidelines_for_database_cross_references#Exact,_broad,_narrow_general_database_cross_references but they could be more precise.

Related cross-references:
I propose that at least cross references that refer to gene products, such as proteases
GO:0008239 dipeptidyl-peptidase activity.
xref: EC:3.4.14.1 : dipeptidyl-peptidase I & MetaCyc:3.4.14.1-RXN
xref: EC:3.4.14.11 : Xaa-Pro dipeptidyl-peptidase
xref: EC:3.4.14.2 : dipeptidyl-peptidase II
(etc), be related synonyms, as these refer to gene products that have different substrate preferences.

Narrow cross-references:

• If the EC or RHEA only differs by the hydrogen donoe/acceptor (NAD versus NADP), then the cross-reference is narrowMatch.
• If the enzyme is specific for several substrates, and there is no single reaction in th eexternal database representing the same enzyme specificity, the cross-reference type is narrow. For example: GO:0047429 'nucleoside triphosphate diphosphatase activity' has these 2 narrowMatches RHEA cross references:
• • RHEA:23996 a ribonucleoside 5'-triphosphate + H2O = a ribonucleoside 5'-phosphate + diphosphate + H+
    and
• • RHEA:44644 a 2'-deoxyribonucleoside 5'-triphosphate + H2O = a 2'-deoxyribonucleoside 5'-phosphate + diphosphate + H+

[pgaudet]

@sjm41 Please add any suggestions. We can discuss this on today's ontology call.

Thanks, Pascale

[sjm41]

It's worth explicitly saying that the EC xref example extends to the MetaCyc xref equivalents.
e.g. Both EC:3.4.14.1 and MetaCyc:3.4.14.1-RXN would be relatedMatch

[pgaudet]

Edited my initial comment to include the MetaCyc example