search

obophenotype / cell-ontology

An ontology of cell types
https://obophenotype.github.io/cell-ontology/
Creative Commons Attribution 4.0 International
142 stars 49 forks source link

[NTR-cxg] BEST4+ intestinal epithelial cell #1695

Closed rachadele closed 1 year ago

rachadele commented 2 years ago

Preferred term label BEST4+ intestinal epithelial cell

Synonyms (add reference(s), please)

Definition (free text, with reference(s), please. PubMed ID format is PMID:XXXXXX) An absorptive epithelial cell of the human intestine expressing Bestrophin 4 calcium-activated ion channels. PMID:24223998 PMID:35176508

Parent cell type term (check the hierarchy here https://www.ebi.ac.uk/ols/ontologies/cl) CL:0000677

What is the anatomical structure that the cell is a part of? (check Uberon: https://www.ebi.ac.uk/ols/ontologies/uberon) intestinal epithelium UBERON:0001277

Your ORCID 0000-0002-7431-4139

Additional notes or concerns If there are markers that uniquely identify the cell type in a particular species, they can be added here.

ghost commented 2 years ago

Hi @rachadele, can you clarify if you intend for this class to only represent a cell type in humans as noted in your definition? I see the reference paper is focused on humans, but if this cell type exists in other species or genera, we can remove the word "human" from the definition and not worry about taxon restrictions.

ghost commented 2 years ago

Hi @rachadele, checking in to see if you have feedback on my comment immediately above. Once clarified, I can move forward with adding this new cell type.

lubianat commented 2 years ago

@bvarner-ebi As far as I know, they are not known in mice.

How about having 2 terms, one more neutral, without the taxonomic restriction, and a subclass restricted to humans?

rachadele commented 2 years ago

Yes, it is intended to be a human cell type! Sorry for the late response.

ghost commented 2 years ago

@bvarner-ebi As far as I know, they are not known in mice.

How about having 2 terms, one more neutral, without the taxonomic restriction, and a subclass restricted to humans?

Thanks for the feedback, @lubianat. My understanding is that it is discouraged to create new classes that would have only one child. (@shawntanzk, is my assumption correct)? Since there is also no evidence supplied that the cell type exists in other species, I am hesitant to create a taxon-agnostic class without supporting dbxrefs.

shawntanzk commented 2 years ago

My understanding is that it is discouraged to create new classes that would have only one child. (@shawntanzk, is my assumption correct)?

Yups. Chris Mungall wrote a blog about this - https://douroucouli.wordpress.com/2022/01/04/ontotip-avoid-the-single-child-anti-pattern/

How about having 2 terms, one more neutral, without the taxonomic restriction, and a subclass restricted to humans?

Would ideally like to avoid this too. My preference with this would be instead to have taxon GCIs rather than a subclass restricted to humans if need be - very easily bloats an ontology if we want everything to have specific terms for species

shawntanzk commented 2 years ago

how about a middle ground where we: 1) remove the human part of the def 2) add 'present in taxon' some 'homo sapiens' 3) add a comment about only being sure that this cell types is in homo sapiens but unsure about other species (from reading the above, it sounds like we are unsure if this is human specific and all)

ghost commented 2 years ago

how about a middle ground where we:

  1. remove the human part of the def
  2. add 'present in taxon' some 'homo sapiens'
  3. add a comment about only being sure that this cell types is in homo sapiens but unsure about other species (from reading the above, it sounds like we are unsure if this is human specific and all)

Thanks for the suggestions. Since the specific cell type requested in this ticket is intended only for humans, my latest commit leaves "human" in the text def, adds the taxon restriction and updates the protein in the axiom to the human form. I did not add a comment since AFAIK that has not been done consistently in the past and would rather not start that precedent.

ghost commented 1 year ago

Hi @rachadele, thank you again for your new term request. The following has been added to CL and will appear in the next release:

CL:4030026 'BEST4+ intestinal epithelial cell, human'

rachadele commented 1 year ago

thank you!

rachadele commented 1 year ago

hi there– when will the release with this term be published?

shawntanzk commented 1 year ago

it should be in the latest release, do let us know if you don't see it and we can track down why :)

rachadele commented 1 year ago

I do not see it!

shawntanzk commented 1 year ago

I've just checked our release (cl.owl at least), it is in there. Perhaps you are trying to find it in OLS? if that's the case, this is due to OLS not having updated CL release - they are working on it, I can't fully rmbr the reason, but their automated system is down, making it tricky. Perhaps consider using ontobee for now - CL purls redirect there anyway (http://purl.obolibrary.org/obo/CL_4030026)

rachadele commented 1 year ago

thank you! I requested this term for a cellxgene submission, but it's causing the submission to fail validation. i'm guessing maybe their validator tool pulls valid ontologies from OLS since it doesn't seem to recognize this term. I will have to ask their devs about it :)

shawntanzk commented 1 year ago

btws just a small thing, but our preference is for people to pull the ontology through our purl system (http://purl.obolibrary.org/obo/cl.owl) - it's something that I'm supposed to tell people using it. Using the purl = if we change anything (eg how we do releases), it won't affect your pipelines. If your devs are going to change where they pull the ontology from, maybe suggest that to them :) thanks!