Add tp53 figure and associated figure organizational changes

sjspielman commented 2 years ago

From Issue #1271, this PR adds a new figure and updates some existing ones. Specifically:

A new figure figures/pdfs/fig3/panels/tp53_panel.pdf and associated legend figures/pdfs/fig3/panels/tp53_legend.pdf
- Instead of using shapes (too indistinguishable) to indicate mutators, I used colors (red and orange for "heat"-like style...)
- Because of how I layered the jitter points to ensure visibility, the legend is separate
The mutational signatures sina plot was made to be horizontal. Currently this plot is copied from its analysis module, hence the diffs for that notebook
Moved GSVA and UMAP figures to be figure 5 instead of figure 4
Renamed the telomerase activities script

New figure scripts were added to generate_figures.sh, and any scripts that hadn't been added in there were also added in.

Directions for reviewers. Tell potential reviewers what kind of feedback you are soliciting.

Should I use the binary palette for hyper/ultra instead? "Normal" is technically a third category.
Be aware that the TMB here is only coding TMB, not all.
This cancer groups in the boxplot are ordered by tp53 score, which will differ from ordering by cancer group size in other panels. I think this plot looks better ordered as is, and think I'd rather see different orderings in plots than change this one's order to match.
I removed a now deprecated overall pdfs/fig4/figure4.pdf and its associated README, in favor of new ones coming down the line. Should we keep these for now?
Did I miss any figure scripts for generate_figures.sh (remembering that Issue #1261 will be coming down the line soon).
As part of this PR I can also add the mutational signatures code into figures/scripts and keep that module as is. Since that figure in the module was created to be pub-ready, I've been using that still.

Is the analysis in a mature enough form that the resulting figure(s) and/or table(s) are ready for review?

yes

Reproducibility Checklist

[x] The dependencies required to run the code in this pull request have been added to the project Dockerfile.
[x] This analysis has been added to continuous integration.

Documentation Checklist

[x] This analysis module has a README and it is up to date.
[x] This analysis is recorded in the table in analyses/README.md and the entry is up to date.
[x] The analytical code is documented and contains comments.

sjspielman commented 2 years ago

@jaclyn-taroni I think the new commits have addressed the legend issue?

jaclyn-taroni commented 2 years ago

Grey background is still there https://github.com/AlexsLemonade/OpenPBTA-analysis/pull/1275#discussion_r838501181

sjspielman commented 2 years ago

Grey background is still there https://github.com/AlexsLemonade/OpenPBTA-analysis/pull/1275#discussion_r838501181 Woops forgot the theme!

sjspielman commented 2 years ago

I updated with theme_pubr() which gives us a horizontal legend by default. Would we prefer vertical?

jaclyn-taroni commented 2 years ago

Would we prefer vertical?

Probably, I think it gives us more flexibility when we assemble things. How about:

+ theme(legend.background = element_blank())

sjspielman commented 2 years ago

No success with + theme(legend.background = element_blank()) or other similar modifications, so I just slapped on a theme_classic(). Seems like it achieves the goal.

jharenza commented 2 years ago

@sjspielman I like the new TP53 boxplots with the hypermutant samples, but wanted to note for organizational purposes, this will be a part of figure 4, not figure 3 (since we don't describe the tp53 scores until figure4).

sjspielman commented 2 years ago

This may need further discussion. Part of this PR flipped the size of the mutational signatures plot to be horizontal specifically so that we could include the hypermutator plot with it. Both plots focus on the same set of cancer groups to explore, and this version of tp53 plot was in part established from an interest to look at hypermutators in conjunction with mutational signatures.

We'll probably want to come to an agreement on this before merging since it may necessitate reverting the mutational signatures plot back to vertical. I don't have very strong thoughts on this, so anyone with strong thoughts please chime in! @jaclyn-taroni @jharenza

jharenza commented 2 years ago

Ahhh, I was initially thinking that for figure 3, we would move the chromothripsis plots alongside the co-occurrence plots, then move mut sigs horizontal as the last panel below. Then, we would transition into figure 4- TP53 scores (original panels ABC), then the distribution highlighting hypermutaters. Then, either faceted below the TP53 distribution, telomerase distribution, followed by a final panel with the hazard plot.

Maybe I should write this up to work on the flow and see if we can get that to work.

jharenza commented 2 years ago

This was what I was sketching

sjspielman commented 2 years ago

Ahhh, I was initially thinking that for figure 3, we would move the chromothripsis plots alongside the co-occurrence plots, then move mut sigs horizontal as the last panel below. Then, we would transition into figure 4- TP53 scores (original panels ABC), then the distribution highlighting hypermutaters. Then, either faceted below the TP53 distribution, telomerase distribution, followed by a final panel with the hazard plot.

This seems good!! I'll move the TP53 boxplots as fig 3 and update READMEs accordingly. I'm working on the other TP53 and telomerase figures before the weekend.

AlexsLemonade / OpenPBTA-analysis