Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[NTR-cxg] Add region neutral children of L2/3-6 intratelencephalic projecting glutamatergic cortical neuron #2154

Closed
jychien opened this issue Sep 19, 2023 · 14 comments · Fixed by #2181

Comments

@jychien
Copy link

jychien commented Sep 19, 2023

Related ticket: #2148

Parent cell type term
The children of L2/3-6 intratelencephalic projecting glutamatergic cortical neuron have the text " of primary motor cortex" at the end of the term labels. To ontologize layer specific neurons that are not definitively from the primary motor cortex, we are requesting that there be terms added that are region neutral.
Additionally, we are requesting that the region neutral grouping terms are added as parents to the layer specific terms:

  • L2/3 intratelencephalic projecting glutamatergic neuron (L2/3 IT) is parent of L2 IT and L3 IT
  • L4/5 intratelencephalic projecting glutamatergic neuron (L4/5 IT) is parent of L4 IT and L5 IT

Preferred term label (Synonym(s) in parenthesis):

  • L2/3 intratelencephalic projecting glutamatergic neuron (L2/3 IT)
  • L2 intratelencephalic projecting glutamatergic neuron (L2 IT)
  • L3 intratelencephalic projecting glutamatergic neuron (L3 IT)
  • L4/5 intratelencephalic projecting glutamatergic neuron (L4/5 IT)
  • L4 intratelencephalic projecting glutamatergic neuron (L4 IT)
  • L5 intratelencephalic projecting glutamatergic neuron (L5 IT)
  • L6 intratelencephalic projecting glutamatergic neuron (L6 IT)

Definition (free text, with reference(s), please. PubMed ID format is PMID:XXXXXX)

  • L2/3 intratelencephalic projecting glutamatergic neuron: An intratelencephalic-projecting glutamatergic neuron with a soma found in cortical layer 2/3. (PMID:37292694, PMID:34004146)
  • L2 intratelencephalic projecting glutamatergic neuron: An intratelencephalic-projecting glutamatergic neuron with a soma found in cortical layer 2 (PMID:34004146)
  • L3 intratelencephalic projecting glutamatergic neuron: An intratelencephalic-projecting glutamatergic neuron with a soma found in cortical layer 3 (PMID:34004146)
  • L4/5 intratelencephalic projecting glutamatergic neuron: An intratelencephalic-projecting glutamatergic with a soma located in cortical layer 4/5. (PMID:34004146)
  • L4 intratelencephalic projecting glutamatergic neuron: An intratelencephalic-projecting glutamatergic neuron with a soma found in cortical layer 5 (PMID:37292694)
  • L5 intratelencephalic projecting glutamatergic neuron: An intratelencephalic-projecting glutamatergic neuron with a soma found in cortical layer 5. (PMID:37292694, PMID:37609206)
  • L6 intratelencephalic projecting glutamatergic neuron: An intratelencephalic-projecting glutamatergic neuron with a soma found in cortical layer 6 . (PMID:37292694, PMID:37609206)

Publications:

Anatomical structure where the cell type is found (check Uberon for anatomical structures:
Cerebral cortex: UBERON:0000956

Your ORCID
https://orcid.org/0000-0003-4389-9821

Additional notes or concerns

  • The parents of the terms containing " of primary motor cortex" would then be the most layer specific of these new terms as possible. For example, "L4/5 IT M1" would be a child of "L4/5 IT" and "L5 IT M1" would be a child of "L5 IT".
  • Was there a particular reason that L2/3 and L4/5 were considered together on not listed as individual layers?
  • NOTE: It would make the most sense that "L5 intratelencephalic projecting glutamatergic neuron of the primary motor cortex" is child of "L5 IT M1".
  • I've made a diagram, since this was gettting so complicated:
Screenshot 2023-10-11 at 2 41 41 PM
@jychien
Copy link
Author

jychien commented Sep 21, 2023

We just received another dataset with the following cell annotations:
['L6-IT-Car3', 'Amy-Exc', 'L6b', 'L6-CT', 'L6-IT', 'L4-IT', 'L5-IT', 'L2/3-IT', 'L5/6-NP', 'L5-ET']

Is it possible to separate L4 with L5?

@ghost
Copy link

ghost commented Sep 22, 2023

Thank you for the additional information, @jychien.

The reference for 'L4/5 intratelencephalic projecting glutamatergic neuron of the primary motor cortex', PMID:33184512, has the following description:

"We were also able to identify a t-type, L4/5 IT_1, that was located on the boundary between L2/3 and L5 and probably corresponds to the quasi-L4 neurons described previously in motor cortex.31"

Given this lack of specificity, the aggregation for L4 and L5 seems reasonable in this case.

Are you proposing to replace L4/5 with L4 and L5 parent classes or to keep L4/5 and add L4 and L5?

@dosumis
Copy link
Contributor

dosumis commented Sep 22, 2023

Any one paper with layer information about a type is likely to focus on the peculiarities of the layer structure of the region being studied. IIRC, the primary motor cortex lacks a clear layer 4 (unlike many other cortical regions), but some the the location of the soma of some the type being discussed in this paper fits with it being quasi-L4 in the PMC.

It is clear that we need a set of region-neutral terms for IT & ET projecting glutamatergic neurons in the cortex. To find out: for which layers can we justify adding ET, IT or both. @jychien - can you provide refs for the papers you're curating for these datasets (or are they pre-release?). Also TBD - consistent system for terms mentioning multiple regions. I think easiest to have L4/5 as a grouping term for neurons in L4 or 5 (so L4 & L5 are child terms) - but it's worth noting that in this case it is referring to a distinct group of neurons that sit between L2/3 and L5.

Please also note that we now have the option to add an official symbol (e.g. L5 ET) and have done so for at least some terms. I'd like to add these for well used symbols to enhance findability and to provide an option for display where long names are not appropriate.

Finally - I think it would be worth having a more general call with folks form the Allen who are working on these cell types to get their take on what's needed. We can't avoid that these names are being used based on annotation transfer in cases were we don't have soma-location and projection info. I can help organise through BICAN. @jychien would you be interested to join??

@jychien
Copy link
Author

jychien commented Sep 22, 2023

Thank you @bvarner-ebi and @dosumis.

To clarify, yes, I was suggesting to keep the L4/5 terms and to add the individual L4 and L5. And it does make a lot of sense to add them as children.

Here are 2 of the preprints that I am working on with contributors:

for which layers can we justify adding ET, IT or both

I have yet to go through all the files, but am starting with IT. And the ET does seem to have a matching term for L5 ET, as you have alluded to. There does seem to be " of primary motor cortex" in other terms, such as L5/6 NP M1, but I didn't want to clutter up this ticket.

Please also note that we now have the option to add an official symbol (e.g. L5 ET) and have done so for at least some terms.

That's great! Is this listed as an exact synonym, or are you saying that there is a new feature of an 'official symbol' that can be retrieved for these cell types with long names?

We do have additional datasets that other wranglers on the Lattice team are working on, and other datasets that are being wrangled from public repositories where we have yet to engage with the authors. It would be great if @jlzamanian and I could be invited to the conversation. Much appreciated! The authors that I am talking with are non-Allen, and it does sound like they have some feedback on the CL terms and is what started this thread. Would it be okay to extend the invitation to these authors, in the case they are interested in adding some feedback?

In general, the datasets for brain that we are getting are large, and if we could get the CL terms in place, it would help alleviate some of the re-annotation work that we would need to do in the future. Thank you!

@ghost
Copy link

ghost commented Sep 26, 2023

To clarify, yes, I was suggesting to keep the L4/5 terms and to add the individual L4 and L5. And it does make a lot of sense to add them as children.

@jychien, thank you for clarifying. Would you kindly update your request to reflect this?

@jychien
Copy link
Author

jychien commented Sep 26, 2023

@bvarner-ebi I've tried to make it as least confusing as possible. Please let me know if something is inconsistent. Thanks!

@dosumis
Copy link
Contributor

dosumis commented Oct 4, 2023

We just received another dataset with the following cell annotations:
['L6-IT-Car3', 'Amy-Exc', 'L6b', 'L6-CT', 'L6-IT', 'L4-IT', 'L5-IT', 'L2/3-IT', 'L5/6-NP', 'L5-ET']

@jychien - is this one of the datasets being processed for loading?

Jorstad, Nikolas L., Jennie Close, Nelson Johansen, Anna Marie Yanny, Eliza R. Barkan, Kyle J. Travaglini, Darren Bertagnolli, et al. 2022. “Transcriptomic Cytoarchitecture Reveals Principles of Human Neocortex Organization.” bioRxiv. https://doi.org/10.1101/2022.11.06.515349.

This is a current priority for our Brain Initiative funding and may become a reference for human cortical types. Looks v. relevant to decisions about cell types being added for this ticket.

@jychien
Copy link
Author

jychien commented Oct 4, 2023

@dosumis Yes, the list of cell annotations is from the second preprint listed above.

Thank you for the link to the preprint. What does it mean for a publication to become a reference for a brain tissue region? Will there need to be any changes to this new terms requested in this ticket? We are looking through our brain datasets in CELLxGENE and do anticipate additional NTRs.

@dosumis
Copy link
Contributor

dosumis commented Oct 5, 2023

What does it mean for a publication to become a reference for a brain tissue region?

It will be added to the list of annotated datasets used as a reference on https://portal.brain-map.org/explore/cell-type-references-and-algorithms. This allows users to upload an h5ad file and get a set of predicted annotations.

@dosumis
Copy link
Contributor

dosumis commented Oct 5, 2023

Please also note that we now have the option to add an official symbol (e.g. L5 ET) and have done so for at least some terms.

That's great! Is this listed as an exact synonym, or are you saying that there is a new feature of an 'official symbol' that can be retrieved for these cell types with long names?

Both. e.g.

https://www.ebi.ac.uk/ols4/ontologies/cl/classes/http%253A%252F%252Fpurl.obolibrary.org%252Fobo%252FCL_4023041

has

image

@dosumis dosumis assigned ghost Oct 9, 2023
@ghost
Copy link

ghost commented Oct 11, 2023

@jychien, thank you for the detailed ticket.
*Note: using abbreviations instead of labels in comments below.

Regarding the placement of L2 IT and L3 IT as subclasses under L2/3 IT, I'm unsure of this modelling.

My understanding is that 'cortical layer II/III' is used when there is no clear distinction between the boundaries of layers II and III as opposed to an arbitrary grouping of the two classes.
Uberon does not classify 'cortical layer II' and 'cortical layer III' as children of 'cortical layer II/III', and it is my inclination to preserve that separation with these new CL classes.

The reference provided for new terms L2 IT and L3 IT also makes a distinction between these layers (emphasis added): "The ENT IT group (Fig. 4BD, ​,FFH) has 10 supertypes organized in a layer-selective manner, in the order of L2, L2/3, L3, L5 and L6, consistent with their correspondence with CTX L2/3-L6 IT types, and shows differential spatial distribution along the anterior-posterior axis..."

The reference also states: "Two supertypes, L2 IT ENTm Lef1 and L2 IT ENTl Chn2 (Grik1+), are located at the border between L1 and L2."
Your proposed text definition for L2 IT has "...soma found in cortical layer 2". Is this distinction in location worth reflecting in the definition (border vs layer)?

With the above in mind, would it be appropriate to classify L2 IT, L3 IT, L4 IT and L5 IT as direct descendants of L2/3-6 IT instead of the hierarchy presently proposed in your request?

CC @patrick-lloyd-ray if you have insight to add.

@jychien
Copy link
Author

jychien commented Oct 11, 2023

@bvarner-ebi Thank you very much for catching that. I hadn't realized the nuanced difference. I've updated the diagram in the first comment accordingly.

@ghost
Copy link

ghost commented Oct 12, 2023

@jychien, thank you for confirming.

For your review, this is how the classes appear in the ontology editor inferred class hierarchy view:

Screenshot 2023-10-12 at 15 53 14

Of note, assuming there was a copy error in the text def for L4 IT, I substituted "4" for "5".

@ghost ghost added the CL-classhierarchy label Oct 12, 2023
@ghost ghost closed this as completed in #2181 Oct 13, 2023
ghost pushed a commit that referenced this issue Oct 13, 2023
…_neuron

Addresses #2154 [NTR-cxg] cortical region-agnostic L2/3-6 intratelencephalic projecting glutamatergic cortical neuron (by layer)
@ghost
Copy link

ghost commented Oct 13, 2023

@jychien, thanks again for the new terms. The following classes have been added and will be available in the next CL release:

CL:4030059 'L2/3 intratelencephalic projecting glutamatergic neuron'
CL:4030060 'L2 intratelencephalic projecting glutamatergic neuron'
CL:4030061 'L3 intratelencephalic projecting glutamatergic neuron'
CL:4030062 'L4/5 intratelencephalic projecting glutamatergic neuron'
CL:4030063 'L4 intratelencephalic projecting glutamatergic neuron'
CL:4030064 'L5 intratelencephalic projecting glutamatergic neuron'
CL:4030065 'L6 intratelencephalic projecting glutamatergic neuron'

Additional details can be viewed in the linked PR.

This issue was closed.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
2 participants