Request the scripts to add custom cell type labels to the Cell Ontology

[Starlitnightly]

Description of feature

Hi,

I noticed the following description in the main text, could you provide the scripts to add custom cell type labels to the Cell Ontology?

Sincerely,
Zehua

"We recommend that in any newly added reference datasets, the annotations should be consistent with Cell Ontology, either by matching terms in the ontology or by hierarchically assigning new terms to existing terms in the ontology. To this end, we provide scripts to add custom cell type labels to the Cell Ontology before processing by popV (where it is used for running OnClass and calculating our consensus scores)."

[shobhitagrawal1]

Thanks to the authors for this repo!
@Starlitnightly the reproducibility repository has some scripts (https://github.com/YosefLab/popv-reproducibility/tree/main/ontology_creation) but I couldn't really figure it out how to do it on my data. Any help would be appreciated, an example notebook would be great.

[canergen]

Hi, thanks for your interest. I hope to get to create a notebook end of this week that walks through adding celltypes and retraining the nlp embedding for OnClass. Currently, as shown in the tutorial, you can set the ontology to false to disable OnClass. If you feel that the OBO ontology is missing important terms, they actively take suggestions in to extend it.

[canergen]

I have pushed a version of it in: https://github.com/YosefLab/popV/blob/cuml_enabled2/add_celltypes_ontology.ipynb. This branch is functional. However, I am retraining the pretrained models for Tabula sapiens still.