OBO term creation date question

[matentzn]

@zhengj2007 @cmungall

We were discussing the inconsistent annotation of terms with either dc:dateTime or oio:creation_date across OBO ontologies. Which is the correction relation we should be using to denote creation time? Whatever it is, I will promote across my community. Thank you!

[matentzn]

Either it seems to be a difficult question or no one knows?

[wdduncan]

I've just been using dcterms:date as an annotation on the term. But oio:creation_date may work better ... more explicit.

[cmungall]

GO uses oio:

http://wiki.geneontology.org/index.php/Protege5_5_setup_for_GO_Eds#Setting_username_and_auto-adding_creation_date

Creator property Add http://www.geneontology.org/formats/oboInOwl#created_by
Creator value Select Use user name
Check Annotate new entities with creation date/time box
Date property Add http://www.geneontology.org/formats/oboInOwl#creation_date
Date value format Select ISO-8601

note that only oio would yield obof tags:

id: GO:0003157
name: endocardium development
namespace: biological_process
def: "The process whose specific outcome is the progression of the endocardium over time, from its formation to the mature structure. The endocardium is an anatomical structure comprised of an endothelium and an extracellular matrix that forms the innermost layer of tissue of the heart, and lines the heart chambers." [GOC:mtg_heart]
is_a: GO:0032502 ! developmental process
is_a: GO:0048856 ! anatomical structure development
intersection_of: GO:0048856 ! anatomical structure development
intersection_of: results_in_development_of UBERON:0002165 ! endocardium
relationship: part_of GO:0007507 ! heart development
created_by: dph
creation_date: 2009-10-01T12:42:27Z

although on option is to translate to oio prior to obof conversion, but IMHO this is awkward and introduces potential confusion

[pbuttigieg]

[not sure if you mean in OBO format alone]

ENVO now defaults to dc terms:

There's a bunch of OBO-specific castings of things like this (including SKOS-like synonyms) which hurts our global interoperability.

[mcourtot]

I like creation_date - as mentioned by @wdduncan above it is less ambiguous; date could be last update date or else for example.

@cmungall what's the status of oboInOwl? I was under the impression that it would be retired and superseded by IAO, OBI or else, but I'm not sure whether this is true. It does seem a little bit messy to have a bunch of resources for what we're hoping to be common annotation properties.

[ddooley]

It would be good to disambiguate creation date from update date. At its simplest dc:date could mean update date, subject to change as edits are made to a term. But then only archival files will allow one to see initial creation date of term.

"dc:date": A point or period of time associated with an event in the lifecycle of the resource

[wdduncan]

Yes. It would be nice to disambiguate between creation date and last modified date. In practice, I don't modify the dc date once the term is created. I track changes using the editor notes. E.g.:

Editor note:
Bill Duncan 2020-01-01:
Made an amazing update to the term's definition!

But this is just the way I've been doing things. Not sure about others.

[cmungall]

This is a dupe, let's use

• 63