Conflicting IRIs for OMIM identifiers

[kshefchek]

Our scigraph data graph has two iri patterns for OMIM identifiers,

Purl version: https://scigraph-data-dev.monarchinitiative.org/scigraph/graph/neighbors.png?id=OBO%3AOMIM_105830&depth=1&blankNodes=false&direction=BOTH&entail=false

OMIM version: https://scigraph-data-dev.monarchinitiative.org/scigraph/graph/neighbors.png?id=OMIM:105830&depth=1&blankNodes=false&direction=BOTH&entail=false&relationshipType=RO:0002200

The consequence is that the mondo graph is not connected to the dipper a-box graph. This means we're missing closures, not categorizing the ids as diseases, and not clique merging.

Potential paths forward:

• Change dipper iris back to purls
• Convert omim ids to purls before they are ingested by scigraph
• Change mondo omim IRIs

This is a blocker for releasing our db.

[TomConlin]

• Have scigraph operate on curies

[cmungall]

Terminological point: when you say "purls" I think you mean "OBO purls"

I think the fix should be made at MONDO. MONDO shouldn't use OBO purls.

For now I think we need to add more logic to https://github.com/monarch-initiative/monarch-ontology builds. We should have more QC checks there, and also inject any temporary hacks in here. We can use robot rename for this.

I don't think we should change scigraph to get around this, and not sure what this would look like anyway, incoming is RDF, standard URIs need to be used.

[nicolevasilevsky]

@kshefchek what is the action item for me?

[kshefchek]

@nicolevasilevsky we need to converge on the IRI mapping for OMIM, we are currently using 'OMIM': 'http://omim.org/entry/'

From https://github.com/monarch-initiative/dipper/issues/700#issuecomment-460366871 it sounds like the action item is to change this in mondo and mint a new release so I'm passing it back to you and @cmungall.

[kshefchek]

The 5/28 version of mondo uses identifiers.org for omim, is there a plan to move to http://omim.org/entry/{id}?

[cmungall]

That shouldn't be the case! Am investigating..

On Mon, Jun 10, 2019 at 8:21 PM Kent Shefchek notifications@github.com wrote:

The 5/28 version of mondo uses identifiers.org for omim, is there a plan to move to http://omim.org/entry/{id}?

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/monarch-initiative/dipper/issues/700?email_source=notifications&email_token=AAAMMOJX4VXTVGR5FDAEGATPZ4K3PA5CNFSM4GUHB662YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODXL2KQA#issuecomment-500671808, or mute the thread https://github.com/notifications/unsubscribe-auth/AAAMMONHJZJHJNZEYRPTGSLPZ4K3PANCNFSM4GUHB66Q .

[cmungall]

Hmm did you mean obolibrary rather than identifier.org?

curl -L -s $OBO/mondo/imports/equivalencies.owl | grep -i omim
...
        <owl:equivalentClass rdf:resource="http://purl.obolibrary.org/obo/OMIM_610706"/>
...

In that case, OK, yes, agreed, we have always used obolibrary which is bad, but this needs to sync up, but at least there was no unexpected change to yet another URI space!

[kshefchek]

this is in the owl file, zgrep -c 'identifiers.org/omim' mondo.owl.gz 8537

[cmungall]

Ah, this is for the skos annotations. These should be in sync for sure. But this shouldn't affect us as we don't use the skos... though I am not totally sure how this ends up looking in scigraph?

On Mon, Jun 10, 2019 at 9:24 PM Kent Shefchek notifications@github.com wrote:

this is in the owl file, zgrep -c 'identifiers.org/omim' mondo.owl.gz 8537

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/monarch-initiative/dipper/issues/700?email_source=notifications&email_token=AAAMMONPGT5CDFUNYI4FUGLPZ4SHFA5CNFSM4GUHB662YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODXL4WAY#issuecomment-500681475, or mute the thread https://github.com/notifications/unsubscribe-auth/AAAMMOPBR74WNGNSEHTM3TDPZ4SHFANCNFSM4GUHB66Q .

[cmungall]

I wrote a piece about the anarchy of mappings... so many ways to represent them https://douroucouli.wordpress.com/2019/05/27/never-mind-the-logix-taming-the-semantic-anarchy-of-mappings-in-ontologie/

[kshefchek]

ah okay I see the obo purls now for OMIM, false alarm! was toying with the mondo-minimal.json and thought the IRIs were replaced

[pnrobinson]

Interesting – your article is getting supplemented by the “Gut doctor: I beg Americans to throw this vegetable out” add. I am beginning to feel like the doctor knows something about me but I cannot seem to figure out which vegetable to discard….or does the add have something to do with OWL? :-0

[monicacecilia]

Is there a status update, @cmungall? This seems to be connected to issues on monarch-initiative/monarch-ui/issues/172

[cmungall]

I have no update not sure if Shahim can help here?

On Wed, Aug 21, 2019 at 6:41 PM Monica Munoz-Torres < notifications@github.com> wrote:

Is there a status update, @cmungall https://github.com/cmungall? This seems to be connected to issues on monarch-initiative/monarch-ui/issues/172 https://github.com/monarch-initiative/monarch-ui/issues/172

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/monarch-initiative/dipper/issues/700?email_source=notifications&email_token=AAAMMONZBMG2GWFTZKYOJDDQFXVCXA5CNFSM4GUHB662YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD43TJ3Y#issuecomment-523711727, or mute the thread https://github.com/notifications/unsubscribe-auth/AAAMMOKFLDDHILO6MJQQKYLQFXVCXANCNFSM4GUHB66Q .

[monicacecilia]

Dear @ShahimEssaid - Do you think you could give us a hand?

[ShahimEssaid]

@kshefchek Let's talk this coming week so I can understand how I can help. I'm not that familiar with the other pieces of this issue.

[kshefchek]

To distill it down, we want to make sure we are using the exact same IRIs in dipper and MONDO, essentially:

In dipper we're currently using OMIM: http://omim.org/entry/1234 OMIMPS: http://www.omim.org/phenotypicSeries/PS1234

As is stands, MONDO uses http://purl.obolibrary.org/obo/OMIM_1234 for skosExactMatch: "http://identifiers.org/omim/1234" OMIMPS: http://purl.obolibrary.org/obo/OMIMPS_1234

We have a terrible hack to normalize the IRIs after building monarch.owl.

[nicolevasilevsky]

@kshefchek is this still an issue?

[nicolevasilevsky]

@kshefchek this ticket is stale, I am closing but please reopen if this is still an issue, thanks!