Count alterations represented in oncoprint on a per cancer group basis

[jaclyn-taroni]

Purpose/implementation Section

We're reporting alterations per cancer group in https://github.com/AlexsLemonade/OpenPBTA-manuscript/pull/199, but we didn't yet have a tabular version of that data. What I'm adding here creates the equivalent tables analyses/oncoprint-landscape/tables for an individual cancer group.

What was your approach?

Here I'm making the following changes:

• Adding a step to analyses/oncoprint-landscape/00-prepare-goi-lists.R where I create a table (analyses/oncoprint-landscape/data/cancer_group_goi_list_mapping.tsv) that contains the mapping between cancer_group and what genes of interest (GOI) list to use. We only create the tables for _all samples in the broad_histology_ when a GOI list is used, so I mirrored that earlier design decision.
• Adding a new script analyses/oncoprint-landscape/04-alteration-counts-by-cancer-group.R which does the following:
  • Reads in MAF, CNV, and fusion files prepped earlier -- in practice, this will be a primary only or primary plus set of files and that's how it works in the shell script right now
  • Uses the table described above to read in the GOI list and create a summary of the MAF that is subset to only the samples in that cancer group and the genes in the GOI list
  • Saves that summary to analyses/oncoprint-landscape/tables/cancer_group_counts where the subdirectory is determined by whether this is for primary only or primary plus (dictated by the shell script)
• Reworks analyses/oncoprint-landscape/03-oncoprint-n-count-table.R to reduce redundant code for primary only vs. primary plus + removes the need to pass broad histology info to that script
• Updates the bash script for the module to reflect these changes, had to make the broad histologies the inner loop.
• Also change to one of the Embryonal oncoprints? To my eye, it only looks like a change in the cancer_group palette. If true, this doesn't matter. The pub-ready plots are made in figures/.

What GitHub issue does your pull request address?

1252

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

Which areas should receive a particularly close look?

• For now, I've only committed the resulting files for DMG otherwise the diff would be 🍌. Please take a close look at those. Or, tell me which cancer group I should use an example instead!
• Does that PNG change look okay?

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

Yes, but note that I am not adding all of the resulting tables yet.

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.

[jharenza]

Thanks @jaclyn-taroni - can you print out either ependymoma or craniopharyngiomas? It looks like I did not separate out DMG within the all HGG oncoprint, but did for these other two.

[jaclyn-taroni]

Yup, added EPN in https://github.com/AlexsLemonade/OpenPBTA-analysis/pull/1266/commits/f714412a41b9edd248fbb37c48e5fa8d9a3217ea

[jaclyn-taroni]

I'm going to mark this ready for review now that I've updated the documentation. I am not going to add all the resultant tables until the code has been reviewed. We're using a maftools function to do the summary, which is to say that I don't really expect (🤞🏻) any edge cases to pop up that wouldn't also pop up when we made the plots.