Update scAdvanced GSEA and AUCell notebooks

[sjspielman]

Closes #898

This PR updates the GSEA and AUCell notebooks for bioconductor 3.22. Changes includes:

• msgidbr now takes the argument(s) (sub)collection and not (sub)category, so I updated this in both notebooks
• For AUCell, I also updated so it no longer uses GeneSetCollections but a named list with gene sets. I also updated the plot_recovery_curve function in utils accordingly. Let me know if there are any parts here you'd like to see done differently (e.g. some purrr code that replaces genesetcollection code)
• I also did some recreational spacing updates while here to clean up a bit.

You can see rendered versions of these notebooks here:

• GSEA: 04-dimension_reduction_scRNA.nb.html
• AUCell: 05-aucell.nb.html

And for comparison, December 2025 schedule links: https://alexslemonade.github.io/2025-december-training/workshop/SCHEDULE.html

There aren't changes to the GSEA output, but there are small changes to some of the AUCell scores. This barely affects thresholds and does not affect results, so these changes are reasonably consistent with what one might expect for a version bump

[sjspielman]

Realizing after filing this that GHA may fail because of similar changes needed to the pathway-analysis notebooks. If/when that happens, I will just have to fold those changes into this PR, but the diffs will be small!

[sjspielman]

Realizing after filing this that GHA may fail because of similar changes needed to the pathway-analysis notebooks. If/when that happens, I will just have to fold those changes into this PR, but the diffs will be small!

Ok, in retrospect this was obviously going to happen.... It didn't fail at the pathway notebooks because it failed sooner, at the cell type notebook which is being fixed in #906. I guess the only way to get CI passing for individual PRs will be to turn it off temporarily (which I don't love the idea of doing - it already caught a bug for us!), or I'll have to make one big PR with all the notebooks.

There are 6 notebooks here with content that needs to change because it errors in bioc 3.22: all the pathway notebooks (2 in scAdvanced and 3 in pathway-analysis; aside from AUCell, this is mostly msigdbr updated syntax), and the singler notebook. Would either @jashapiro or @allyhawkins be up for reviewing a larger PR, or would like to suggest a different strategy?

[jashapiro]

Would either @jashapiro or @allyhawkins be up for reviewing a larger PR, or would like to suggest a different strategy?

You can temporarily disable rendering for notebooks in https://github.com/AlexsLemonade/training-modules/blob/master/scripts/render-live.sh

If you do that, I would file a separate issue just to be sure everything in reenabled as a final test (but try to reenable as you go)

[sjspielman]

You can temporarily disable rendering for notebooks in https://github.com/AlexsLemonade/training-modules/blob/master/scripts/render-live.sh

Just to be extra sure - I read this as suggesting commenting out lines and then restoring as I go, yes?

[jashapiro]

Just to be extra sure - I read this as suggesting commenting out lines and then restoring as I go, yes?

Yes, correct.

[allyhawkins]

LGTM

[allyhawkins]

I was a little on the fence on removing this because the GeneSetCollection is still supported, but I think I agree that this is cleaner and more straightforward.