Ingest DO, ORDO, OMIM, NCIT mappings

[matentzn]

• Have separate goals for obtaining the sources, which may eventually be extended to contain preprocessing (e.g. mirror/do.owl:)
• Using sssom-py extract all mappings from DO, ORDO, OMIM and NCIT.
• Manually review a dozen or so in the source ontology whether they don't contain extract curation information (ORDO ntbt for example)

[matentzn]

• Can you double-check that ncit.owl contains the ICD 10 mappings mentioned here:
  https://ncit.nci.nih.gov/ncitbrowser/pages/mapping_search.jsf?nav_type=mappings

[hrshdhgd]

You mean NCIt_Maps_To_ICD10_2016 (22.01e) and NCIt_Maps_To_ICD10CM_2017 (22.01e) ?

[matentzn]

Yeah, maybe.. Not sure. I guess I am wondering whether ncit.owl contains the mappings mentioned on this page, and if the page is mentioning these two, then yes.

I would love these mappings, especially the newest one, to be included here!

[hrshdhgd]

So ncit.owl has just the mention of ICD10 2017 in them. Does not seem to have the actual ICD10 code. Although the mapping tables linked above have the mapping: 800 (ICD10 2017), 15 (ICD10 2016) and have the corresponding NCIT code.

[matentzn]

Perfect, thank you for checking. In this case, let's build an ingest into SSSOM for these!

[hrshdhgd]

So do you mean having a new method in sssom-py that accommodates for external mapping files?

[matentzn]

Bit unsure where to draw the line. Maybe for now a simple python script that imports SSSOM mapping and mapping sets and just do it externally? I told @joeflack4 to add the ICD10-snomed mapping parser to SSSOM py but I am not sure. Maybe yes, adding it to parsers would be an option. For now the answer is: python, in whatever way you like. But perhaps having a method from_ncit() in parsers.py would be awesome.

[hrshdhgd]

Alright, for the time being I'll just do an external JOIN so to speak to include it in the SSSOM mappings as a quick-n-dirty solution. We can discuss a generic solution down the road after having experience with the different scenarios.

[matentzn]

Sounds good!

[hrshdhgd]

ICD mapping column:

RELATIONSHIP_TO_TARGET

Broader Than
Related To
Has Synonym
Narrower Than

Which one would be appropriate?

"Broader Than": "skos:broader",
"Related To": "skos:relatedMatch",
"Has Synonym": "skos:altLabel",
"Narrower Than": "skos:related",

or

"Broader Than": "oboInOwl:hasBroadSynonym",
"Related To": "oboInOwl:hasRelatedSynonym",
"Has Synonym": "oboInOwl:hasExactSynonym",
"Narrower Than": "oboInOwl:hasNarrowSynonym",

[matentzn]

It's confusing but bare with me:

• Broader than: skos:narrowMatch
• Narrower than skos:broadMatch

[matentzn]

Related to : skos:relatedMatch

We don't need synonym unless the object of synonym is a code and not a label. What about "equivalent"?

[matentzn]

Maybe synonym is equivalent

[hrshdhgd]

• skos:closeMatch,
• skos:exactMatch,
• owl:sameAs,
• owl:equivalentClass,
• owl:equivalentProperty

from https://www.w3.org/TR/skos-reference/#L4858

[matentzn]

Use skos:exactMatch in these cases

[matentzn]

Just for recording (no need to concern yourself): NCIT versions can be found here: https://stars.renci.org/var/NCIt/releases/)

[matentzn]

https://thesaurus.cancer.gov/ncitbrowser/pages/mapping_search.jsf?nav_type=mappings&b=0&m=0