Repertoire integration with MuData?

[crichgriffin]

Hi,

I was just wondering whether you have put in any thought to how repertoire data might be included in muon?

I guess you would support scirpy https://scverse.org/scirpy/latest/index.html since it is the only main option in the scverse.
When you read in data with scirpy it creates an anndata object which is empty apart from the obs, which you can choose to merge with another anndata assay such as rna.

The question is how to incorporate the data? Does it get stored in the mdata.obs for the mudata object, or as an anndata within a separate modality e.g. mdata['tcr'].obs . The second would be my preference, to keep it separate, especially if you had TCR and BCR data that you wanted to process separately. But storing it in the mdata obs is more similar to how scirpy works.

Anyway, I would welcome your thoughts on the matter

Charlotte

[gtca]

Just to leave is as a reference for those who stumble upon this discussion: it happens in the respective issue scverse/scirpy#327.
(And thanks for bringing this up, @crichgriffin!)