search

geneontology / minerva

BSD 3-Clause "New" or "Revised" License
6 stars 8 forks source link

Use new properties (RO, IAO) for evidence and model PMIDs as individuals #13

Open hdietze opened 9 years ago

hdietze commented 9 years ago

We need to update the default properties for evidence handling:

The requires that IAO is added to the import chain. Also verify first the RO:0002612 is in the current tbox.

TODO:

hdietze commented 9 years ago

Coordinate with @jnguyenx and @kltm synchronized changes (JS libraries)

kltm commented 9 years ago

I'll be folding these changes in as I work on geneontology/noctua#141 .

hdietze commented 9 years ago

Problem: RO:0002162 is not axiom has evidence

hdietze commented 9 years ago

The correct relationship id is: http://purl.obolibrary.org/obo/RO_0002612 or RO:0002612 Please note the reversal of 6 and 1 here.

hdietze commented 9 years ago

Also we need to decide on the relation (and type) for the with instances. At the moment this is modeled as a string literal in the evidence object

cmungall commented 9 years ago

Note that is-about goes from publication to evidence. We can define an inverse if preferred

cmungall commented 9 years ago

Documentation here:

https://github.com/geneontology/minerva/blob/master/specs/owl-model.md#axiom-annotations-and-evidence

schema: diagram

cc @mchibucos

kltm commented 8 years ago

@cmungall you might want to ditch the old diagram and/or add the new one. Or a ref to the most recent.

kltm commented 8 years ago

Clarification here about the virtual edge's "axiom has evidence": that's just what we're pretending in our heads right? In reality (i.e. file, client model) we're just pointing to an individual by id reference...

I see a conversion on the board for this, from the made-up evidence IRI to RO:0002162: https://github.com/geneontology/noctua/issues/141#issuecomment-125382961 ; is that still in the cards?

cmungall commented 8 years ago

Correct. An OWL annotation value cannot be an Object, it can only be the IRI of the object. I debated whether to explicitly represent this. When the diagram is presented alongside the example OWL this should hopefully be clearer: https://github.com/geneontology/minerva/blob/master/specs/owl-model.md#axiom-annotations-and-evidence

This distinction (IRI != Object) only happens at the OWL interpretation level. The diagram is a precise representation of the RDF, where IRI == Object, and the RO:0002162 triple points directly at the evidence individual.

is that still in the cards?

not sure I understand the question

cmungall commented 7 years ago

This ticket could be split