HumanCellAtlas / ontology

3 stars 1 forks source link

[ENQ] What's the best Uberon term to associate to iPSCs? #105

Closed paolaroncaglia closed 2 years ago

paolaroncaglia commented 2 years ago

From a Slack thread between @pnejad and myself:

"Sometimes we have induced pluripotent stem cells as our cell lines and we need to list a model organ for them from uberon. What would be an ‘organ’ for iPSCs?"

"If the dataset, publication, or cell line "manufacturer" don't provide detailed info on how exactly the iPSCs were derived, you could go with UBERON:0007023 'adult organism'. Based on the knowledge that iPSCs are generated from somatic cells coming from adult individuals/patients. I know it's a broad term, but if you need to list an anatomy term just for technical reasons, would that work? If not, I'd be happy to look at examples if you could send them to me please."

"We’ve been using embryo as the model organ for ipsc, since the pluripotent state is similar to early embryonic state. I feel like adult organism would imply that the differentiation has already occurred. Is there something we could use in place of embryo that could represent the pluripotent state?"

"It'd be helpful if you could please point me to a couple of public datasets/publications where I can look at the Materials & Methods sections. Also please to confirm, do you mandatorily have to use a term from Uberon? (Due to metadata schema etc.)"

"The hca metadata schema requires that we list model organs/the term to be from uberon for anatomy terms. Example of a dataset with iPSCs that were cell lines: https://www.nature.com/articles/s41467-021-25328-6#Sec10 Another dataset that has been submitted to the HCA: https://stemcellsjournals.onlinelibrary.wiley.com/doi/full/10.1002/stem.3302"

paolaroncaglia commented 2 years ago

Both publications provided above send into a loop of looking up older references to see what organ/cell types the iPSCs were originally derived from (and note, if it's a cell type that won't work for HCA).

In both cases above, I've seen mention of fibroblasts and blood. This summary from a UCLA resource is also in agreement: "iPSC are derived from skin or blood cells that have been reprogrammed back into an embryonic-like pluripotent state..."

So if the derivation of iPSCs used in an experiment can be traced back to either skin or blood, the Uberon terms for 'skin of body' or 'blood' may be used. These terms don't specify a development stage, so may in principle be used for embryo as well as adult.

If the derivation of iPSCs isn't easily traceable, unfortunately the common ancestor of skin and blood is very high up in the ontology tree (material anatomical entity, as blood is not considered an organ). So you may as well avoid specifying an organ, and go with the Uberon terms for 'embryo' or 'adult organism' - these are still, formally, anatomy terms.

In this respect, I can see the rationale behind preferring 'embryo' ("We’ve been using embryo as the model organ for ipsc, since the pluripotent state is similar to early embryonic state. I feel like adult organism would imply that the differentiation has already occurred"). But, as far as I've read, iPSCs are always derived from adult tissue or cells, and then "triggered back" into a pluripotent state. If you are able to use a cell type term in addition to the mandatory anatomy one, I'd suggest a combination of UBERON:0007023 'adult organism' + CL:0002248 'pluripotent stem cell' to capture the pluripotent state. Would this work for you?

paolaroncaglia commented 2 years ago

@pnejad Please see my comment above about iPSCs. Thanks.

pnejad commented 2 years ago

Hi @paolaroncaglia, Thank you for looking into this. This works for me! I'll run it by the rest of the team to get their feedback as well.

paolaroncaglia commented 2 years ago

@pnejad

This works for me! I'll run it by the rest of the team to get their feedback as well.

Great, thanks. I'll leave this ticket open for a few days, so if there are any concerns, please feel free to add them here.

paolaroncaglia commented 2 years ago

@pnejad I'll close this ticket now, but if your team had any concern with my suggestion on how to represent iPSCs, please feel free to re-open. Thank you.