Figure generation script + other repo cleanup

[jjc2718]

This PR might look like a lot of changes in terms of lines, but most of them are just cosmetic changes to figure generation and other small details.

Major changes:

• Added a script to generate publication figures (generate_figures.ipynb)
• Renamed scripts in 02_classify_mutations to correspond more closely to which parts of the paper they're relevant to
• Changed how we're doing statistical testing slightly:
  • Before we were doing unpaired, equal-variance t-tests for all of our statistical comparisons (then correcting for multiple comparisons using FDR correction)
  • Now we're doing paired t-tests (since CV folds will have the same train/test labels, so fold 1, model 1 will be paired with fold 1, model 2, etc).

These aren't yet the final versions of the figures we'll use for the paper (I still have to finish running some things and update the figures), but this PR will put in place the infrastructure for regenerating the publication figures once the data is generated.

[jjc2718]

Nice nice figures. @danich1 is a trend setter!

Yeah, svgutils is so much easier than having to remake Inkscape figures every time I change a panel! Super cool stuff.

Some comments on the figures as you finalize them, so these changes don't need to be made in this PR:

• Panels and text for title and axis are fairly small
• (super nitpicky) Not sure if you want to change the colors for your "volcano" plots since the same colors are used in the box plots but they aren't associated with each other
• Looks like you change colors for your data type in figure 6

These are all great points. The figures are definitely not finalized yet (I still have to tweak/rerun some things) but I'll keep these in mind as I remake them.

Created #50 to remind myself.