Gregor Sturm

> What would you do in the instances of cells not having AIRR data? fill the corresponding rows with `NaN`s? This is how scirpy also handles this currently. Although this...

That's a fair point, I need to try this out, but it should be very fast. Awkward array claims to be similarly scalable as numpy arrays.

### where to store the data (`.obs` vs. `.obsm` vs. `.X`) > The AIRR assay data (equivalent to .X) is stored as a [BumpyMatrix](https://bioconductor.org/packages/release/bioc/html/BumpyMatrix.html), specifically a BumpyDataFrameMatrix. We have a...

@javh in R, did you use any specific package for immune receptor analysis?

> @grst, I use the [Immcantation](https://immcantation.readthedocs.io/) packages, because I'm biased like that. :) I'm asking because I was wondering if we can extend this effort beyond the scverse world and...

After the input from @javh and the discussion in https://github.com/scverse/anndata/pull/647, it seems cleanest to me to go the `mudata` route with AIRR data stored in `.X`. Prerequisite is that anndata...

> For what it's worth, on the R side we've been using [MultiAssayExperiments](https://bioconductor.org/packages/release/bioc/html/MultiAssayExperiment.html) with a "rearrangement" experiment that includes the AIRR Rearrangement data for storing multimodal single-cell data that includes...

In GitLab by @szabogtamas on Apr 6, 2020, 10:53 created merge request !40 to address this issue

In GitLab by @szabogtamas on Apr 6, 2020, 10:53 mentioned in merge request !40

In GitLab by @grst on Apr 6, 2020, 14:35 changed the description