Add subset tag for CellxGene usage

[dosumis]

Investigate whether we can mine CxG Census to find all CL terms in use on CxG. If yes, build GA pipeline to mine CxG for a CL term list + ROBOT pipeline to tage CL terms with subset tag: cellxgene_subset. (If not possible - request functionality from CxG Census?)

@anitacaron - this is very similar to what we already do for HRA. Can you point @hkir-dev to existing pipelines?

[anitacaron]

The pipeline for HRA is in the ccf-validation-tools. It might be better to use the subset pipeline from ODK, the same as the other subsets here, e.g. eye, kidney, etc.

[anitacaron]

Or generate a component from the ROBOT template.

[hkir-dev]

I prepared an experimental project https://github.com/hkir-dev/cellxgene-cell-reporter

A nightly build GitHub action is generating

• A robot template: https://github.com/hkir-dev/cellxgene-cell-reporter/blob/main/templates/cellxgene_subset.tsv
• An ontology: https://github.com/hkir-dev/cellxgene-cell-reporter/blob/main/cellxgene_subset.owl

@anitacaron how can we integrate this to CL?

[anitacaron]

• 1. Update the cl-odk.yaml following this example:

  cell-ontology/src/ontology/cl-odk.yaml

  Lines 113 to 116 in bf0391d

  	- filename: kidney_upper_slim.owl
  	use_template: True
  	templates:
  	- kidney_upper_slim.csv

The template file should be in src/templates. ODK will generate the owl file in the src/ontology/components, no need to add it.

• 2. Inside src/ontology, run sh run.sh make update_repo

• 3. Add the component URL to the catalog: example

  cell-ontology/src/ontology/catalog-v001.xml

  Line 21 in bf0391d

  <uri name="http://purl.obolibrary.org/obo/cl/components/kidney_upper_slim.owl" uri="components/kidney_upper_slim.owl"/>

• 4. Add the component URL as an import in the cl-edit (in alphabetical order): example Import(<http://purl.obolibrary.org/obo/cl/components/kidney_upper_slim.owl>)

• 5. Add a rule to update the template in cl.Makefile: the example is downloading a component because there isn't a template, but in your case, you should download the template. (In the next ODK version, this step won't be necessary; we'll be able to just pass the link to an external template in the cl-odk.yaml file)
  

  cell-ontology/src/ontology/cl.Makefile

  Lines 230 to 235 in bf0391d

  	HRA_SUBSET_URL="https://raw.githubusercontent.com/hubmapconsortium/ccf-validation-tools/master/owl/CL_ASCTB_subset.owl"
  	$(TMPDIR)/hra_subset.owl:
  	wget $(HRA_SUBSET_URL) -O $@
  	$(COMPONENTSDIR)/hra_subset.owl: $(TMPDIR)/hra_subset.owl
  	$(ROBOT) merge -i $< annotate --ontology-iri $(ONTBASE)/$@ --output $@

• 6. Inside src/ontology, run sh run.sh make components/{owl file name} to generate the owl file

Let me know if you need further help.