Update cell type notebook for Bioconductor 3.22

[sjspielman]

Closes #897

This PR updates the cell type notebook for some updated SingleR behavior.
It has a new argument test.genes which needs to be provided if the genes in the reference and query are not the same in the same order. The restrict argument controls which genes are included in the DE, and isn't needed now that we have to supply test.genes. From the current version's docs:

Character vector of the names of the genes in the test dataset, i.e., the row names of test in classifySingleR. 
If NULL, it is assumed that the test dataset and ref have the same genes in the same row order.

I also made some recreational spacing updates while I was here.

Kind of annoyingly, this update doesn't appear to be reflected in rdrr which we link in the cheatsheet, but it is in the current version of the singler book.

Here is a rendered version of this updated notebook: 06-celltype_annotation.nb.html
For comparison, here is the rendered notebook at bioconductor 3.19; results are consistent! https://alexslemonade.github.io/2025-march-training/completed-notebooks/scRNA-seq/06-celltype_annotation.nb.html

[sjspielman]

This GHA failed at an earlier notebook that I did not previously have any issues running in the Docker image: https://github.com/AlexsLemonade/training-modules/actions/runs/21910222166/job/63261677007?pr=906#step:6:25

Will investigate and try to address (possibly separate PR) before marking this ready for review.

[sjspielman]

This is now hitting an error at the AUCell notebook, which actually doesn't surprise me. But, we are past the bug that originally caused the GHA here to fail!

Because of this, I'm going to put this PR on pause and tackle #898 first, which will circumvent the error here because we're switching away from genesetcollections -> lists of marker genes.

[sjspielman]

Looks like we've got enough space for now at least! re-#912

[jashapiro]

LGTM