Create Panels for Figure S5

[jharenza]

Three panels for figure S5 (see here) need to be created and colors updated to be consistent with others.

A: https://github.com/AlexsLemonade/OpenPBTA-analysis/blob/master/analyses/tp53_nf1_score/plots/polya_TP53_roc.png B: I removed B (https://github.com/AlexsLemonade/OpenPBTA-analysis/blob/master/analyses/tp53_nf1_score/plots/tp53_expression_by_altered_status_polya.png) because I don't think it is needed - we don't need to make the activated/loss comparison with the classifier not working on these samples. Instead, we should have had a plot of activated/loss/other, but perhaps all samples were activated/loss. Still, don't think we need to go beyond showing the ROC. C: EXTEND vs TERT expression: https://github.com/AlexsLemonade/OpenPBTA-analysis/blob/master/analyses/telomerase-activity-prediction/03-Comparing-TERTexp-TERCexp-EXTENDScores.R D: EXTEND vs TERC expression: https://github.com/AlexsLemonade/OpenPBTA-analysis/blob/master/analyses/telomerase-activity-prediction/03-Comparing-TERTexp-TERCexp-EXTENDScores.R

Looks like C and D are not being saved, and currently, we were showing plots for the stranded samples, although for the boxplot figures (and survival PR here), we use all samples because both polyA and stranded performed well. So, should we lump all together or just show stranded and mention polyA is similar? I suppose we could also do a multi-cor plot similar to S2-A if we want to show all of them. I can go either way and we can explain - I don't think we need to show all, but we can.

Who will complete this? @sjspielman ?

[sjspielman]

Noting this is partly covered in Issue #1272

[sjspielman]

Remaining have been completed and merged in #1283. Some panel re-organization has occurred since this issue was filed, so at time time all of Figure S5 is complete.