Handle NOT qualifiers in GPAD relation column
Closed this issue · 12 comments
Figure out where these NOT qualifiers go in the model. I'm guessing an axiom is involved. An example GPAD line from wb.gpad:
WB WBGene00000038 NOT|enables GO:0003990 PMID:10891266|WB_REF:WBPaper00004251 ECO:0000320 20170131 WB
We still need to figure out how to import 'NOT' qualifiers. For example, there should be a 'NOT' qualifier on the 'heme binding' MF annotation in this model:
http://noctua-dev.berkeleybop.org/editor/graph/gomodel:510042ce-d636-412f-aee4-764478bac44a
The OWL expression for NOT is 'complement of'. e.g.
In the turtle it would look like this:
http://model.geneontology.org/510042ce-d636-412f-aee4-764478bac44a/f6d4e593-0907-4d88-ac6a-0af59d62fd80 rdf:type owl:NamedIndividual , [ rdf:type owl:Class ; owl:complementOf http://purl.obolibrary.org/obo/GO_0020037 ] ;
You need to figure out how to make ontobio build an OWL class expression and then link the instance to that expression via rdf:type.
You need to figure out how to make ontobio build an OWL class expression and then link the instance to that expression via rdf:type
This would be done via rdfslib and should be straightforward, @dougli1sqrd can help
Will PR the change in ontobio under biolink/ontobio#286 and use the new ontobio function.
Just to open a weird can of worms here, how does one write NOT in gpad? What are the rules? http://geneontology.org/docs/gene-product-association-data-gpad-format/#qualifier does not talk about NOT. Should we update the documentation?
An example line w/ NOT:
MGI MGI:2385656 NOT|part_of GO:0005777 MGI:MGI:3687083|PMID:14516277 ECO:0000314 20091006 MGI
@dustine32
What is the current status of importing annotations with a NOT qualifier? Is there anything outstanding wrt the ShEx?
Just looking at an MGI gene with NOT annotations....
In AmiGO, I see this for MGI:MGI:2385656:
but the corresponding GO-CAM model doesn't yet show the corresponding NOT annotations:
Thx.
@vanaukenk Right, the NOT translation has yet to be implemented so it wouldn't appear in the noctua-dev models. I can bump this up in priority since it's likely a simple thing to code and then we can test against ShEx.
Okay, that sounds great. Thanks @dustine32
We can prioritize this for the imports after this week's meeting.
@dustine32
Do you think we're finished with this one? If so, please feel free to close. Thx!
@vanaukenk Yeah, this one looks to be finished. Thanks!