Zenodo CSV: Figure 5 and part of S6

[sjspielman]

Towards #1692 ⚠️ Stacked on #1707

This PR adds CSVs for all figure 5 panels, and the non-UMAP figure S6 panels, because they are in one of the figure 5 scripts.

5A & 5B: figures/scripts/fig5-panels-gsva-umap.R

• 5A is the UMAP. This was pretty straight forward, and you might like some of the piping changes. One change I made was removing the line select(X1, X2, broad_histology_display), since this ends up removing sample ID information which we'd like to keep in the export, and it doesn't hurt to be there during plotting.
• 5B is the GSVA heatmap. I decided to (but feel 50/50 about it) transpose the data for export so that the biospecimen ID would be a column itself instead of column names. The exported (wide) CSV therefore has pathways as column names.

5D: figures/scripts/fig5-forest-plot.R

This is another forest plot. You'll be happy to see the code is about the same for all forest plots in the paper! I exported the non-spread data frame.

5C, 5E, S6E, S6F: figures/scripts/fig5_figS6-immune-deconv-panels.R

These are all pretty straightforward plots, nothing too interesting about these during export. In the end, the code was all roughly the same for each of the 4 exports, so I suppose I could have written a little wrapper function, but I did not...

[sjspielman]

UMAP plot (and other) outputs including hex values seems a bit excessive, but I'm not sure it is worth removing.

In general, I think it needs to be there when we've used* scale_color/fill_identity(), but otherwise it's probably just noise (although harmless!).