search

HumanCellAtlas / ontology

3 stars 1 forks source link

[ENQ] What would be the correct ontology to use for mouse strain = NIH/Swiss? #79

Closed pnejad closed 3 years ago

pnejad commented 3 years ago

Our dataset used NIH/3T3 cells that came from an NIH/Swiss strain embryo. What would be the correct EFO ontology to use here? https://www.atcc.org/products/all/crl-1658.aspx

paolaroncaglia commented 3 years ago

Hi @pnejad,

The right term to use would be EFO:0001222 NIH3T3. It refers the same ATCC entry that you listed, but I’m guessing you couldn’t find it in EFO because of the different wording (NIH3T3 instead of NIH/3T3).

@zoependlington @dosumis it seems that cell lines (i.e. descendants of CL:0000010 ’cultured cell’) are not imported in HCAO? They surely don’t have FMA xrefs, and by looking at their path up the EFO graph and at the efo_import and efo_slim files in https://github.com/ebi-ait/ontology/blob/master/src/ontology/, it seems to me that they wouldn’t be imported automatically. I can’t see cell lines in the HCA OLS production instance. Have HCA wranglers not needed cell line terms before? If so, what’s the best strategy: a) can we import EFO:0001222 NIH3T3 alone, or b) do we have a good enough use case to justify importing the whole ‘cultured cell’ branch which is huge.

Meanwhile, I will edit EFO:0001222 NIH3T3 to add

(now https://github.com/EBISPOT/efo/issues/1047) (done, the changes will be in the EFO release scheduled for April 15th)

@pnejad Lastly, EFO:0001222 NIH3T3 is already linked to the anatomy term for “embryo”. It doesn’t refer the NIH/Swiss strain that the cell line came from, and that you mention, but the information can be gained from the ATCC reference if need be, so I wouldn’t worry about that.

Thanks, Paola

pnejad commented 3 years ago

@paolaroncaglia Thank you for the detailed explanation!

paolaroncaglia commented 3 years ago

Hi @pnejad ,

For your request to annotate NIH/3T3 cells, @zoependlington imported CLO:0008172 (NIH-3T3 cell) into HCAO today. It will be visible in the HCA OLS production instance as soon as this ticket is addressed.

Going forward, if/when you need more cell line terms, ideally we'd want to reference the domain ontology, CLO, if the term is available there. You may look terms up in https://www.ebi.ac.uk/ols/ontologies/clo and let us know that you'd like to import them - FYI, Zoe has set up the makefile for HCAO in that sense.

Thanks, Paola