A little advice

[libmelo]

Hi, thank you for developing this useful tool.
There are two things I want to report.

1. I encountered an error and found that it was because adata.X were "stored elements in Compressed Sparse Row format" sometimes. And at this time, cello.scanpy_cello can not work. (It can be solved by adata.X.toarray())
2. I found cello.scanpy_cello has a useful parameter remove_anatomical_subterms. And I think perhaps we want to limit the cell types (eg. lung, liver or breast) other than to filter out some other types in general?

[mbernste]

Hi, thank you very much for your feedback! Regarding 1.) yes indeed CellO does not work on sparse matrix formats. That is a feature we would definitely like to add. 2.) Can you clarify what you mean by this point? This parameter enables you to provide an ontology term from the Uberon ontology describing anatomical entities, and it will filter out cell type terms from the results that are annotated to be a "part of" the listed anatomical entity.

[libmelo]

For 2), I mean in general we know the types of our samples and want to restrict the cell types for annotation. For example, I may got an annotation of 'glial cell' for breast cancer cells, which make me comfused.
Maybe it's more effective by restricting the types.
Besides limited by my knoledge, I can't use the parameter with Uberon ontology to filter out some types(glial cell and some others).