search

monarch-initiative / mondo

Mondo Disease Ontology
http://obofoundry.org/ontology/mondo
Creative Commons Attribution 4.0 International
223 stars 51 forks source link

Request to include xrefs from Mondo to ClinGen #6595

Open twhetzel opened 10 months ago

twhetzel commented 10 months ago

Following up from the Mondo Outreach call presentation on ClinGen (25-Aug-2023), a request was made to include xrefs from Mondo back to ClinGen to better support users in finding clinical evidence about a disease. This ticket is a general placeholder ticket to capture this request and more details will be added as this task is further defined.

matentzn commented 8 months ago

This is a great move. If you could coordinate with ClinGen to provide us with a two column spreadsheet with MONDO_ID and ClinGen Xref, and some way URL that can be used to resolve the ClinGen xref to their website, we can do this quickly!

twhetzel commented 8 months ago

Email sent to request the information.

matentzn commented 7 months ago

Maybe have an epic to standardise our xref (not mapping!) pipeline? I have open issues for updating Medgen, NORD, ClinGen, GARD xrefs at the very least - this would be a great feature for the December release of Mondo

nicolevasilevsky commented 7 months ago

@twhetzel I can help you with this, if needed

twhetzel commented 7 months ago

@nicolevasilevsky sure, that would be great! I don't have this file yet though and will send a follow-up email.

@matentzn sounds good. How is this group managing "epics" in GitHub?

matentzn commented 7 months ago

@matentzn sounds good. How is this group managing "epics" in GitHub?

not.

I personally fake epics by adding checklist elements in the first comment of the issue, which are interpreted by Github similar as "dependent issues" are interpreted by ZenHub. We never had the need to really up our game on Epics, but - maybe we should (as always, in small incremental steps, ideally not by switching the whole org to ZenHub as the "first step", unless this cant be avoided)!

twhetzel commented 7 months ago

The file is in the Mondo shared drive in Community provided files.

matentzn commented 7 months ago

The file you shared results in around 2000 assertions that look like this:

[Term]
id: MONDO:0000030
name: sleep-related hypermotor epilepsy
subset: clingen {source="MONDO:CLINGEN"}
synonym: "epilepsy, nocturnal frontal lobe" EXACT [OMIMPS:600513]
synonym: "sleep-related hypermotor epilepsy" EXACT CLINGEN_PREFERRED [PMID:27164717]
xref: CGGV:assertion_0f832b8a-af5a-494e-bfdb-70cc619addf1-2020-12-25T015652.186Z {source="MONDO:CLINGEN"}
xref: CGGV:assertion_e3aa0108-5514-48ba-a09c-5506b2030d1f-2020-02-18T170000.000Z {source="MONDO:CLINGEN"}
xref: CGGV:assertion_ed365ac2-8cfa-4ae9-8fef-323dd596a22f-2023-02-07T170000.000Z {source="MONDO:CLINGEN"}
xref: OMIMPS:600513 {source="MONDO:equivalentTo"}
is_a: MONDO:0002612 ! frontal lobe epilepsy
relationship: has_characteristic MONDO:0021152 {source="OMIMPS:600513"} ! inherited
property_value: IAO:0000233 "https://github.com/monarch-initiative/mondo/issues/3891" xsd:anyURI

These seems to link individual CLINGEN assertions to ~1300 Mondo ids.

Wouldn't it be better to instead do a single assertion per MONDO ID:

xref: clingen.condition:MONDO:0020119 {source="MONDO:CLINGEN"}

It also seems that ClinGen does not have its own disease identifiers, and uses Mondo instead, hence the unusual xref clingen.condition:MONDO:0020119.

matentzn commented 7 months ago

I will hold off on this until we had a chance to talk what is best.

EDIT: We could add these as "rdfs:seeAlso" as well, instead of "cross-references".

matentzn commented 1 week ago

We now have this new property: https://github.com/monarch-initiative/mondo/pull/7792

We can move this forward!