Lancet exploration notebook series: Part 1 TCGA-PBTA comparison variations

[cansavvy]

Purpose/implementation Section

What scientific question is your analysis addressing?

This first notebook of three was an exploration into how different callers were resulting in different PBTA- TCGA comparisons.

There are three questions and plots to accompany them :

1. Is the read depth different between PBTA and TCGA
2. Do TMB comparisons results change if we calculate TMB with each caller by itself?
3. How much do the TCGA and PBTA overlap in their target WXS regions?

What was your approach?

Respectively:

1. Plot the read densities for TCGA and PBTA
2. Recreate the TMB comparison CDF plot for each caller
3. Create Venndiagrams of the overlaps of the target regions

What GitHub issue does your pull request address?

#548

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

Two headlining notes before you get into the nitty gritty:

1. This PR is huge. Let me know if you want me to split it up.
2. Note that this analysis is gonna be retired almost as soon as we get it in (and its already outdated with the BED file changes that are still in progress Updated analysis: bedtools creation of bed files for TMB #564 and Bed files intersection fixes #566 )

Which areas should receive a particularly close look?

There are some parts of this analysis that could be DRY'ed up, but it also will be retired as soon as we get it confirmed. Let me know what parts of this code clearly need work and which we will leave as is.

Results

What is your summary of the results?

Here's the rendered notebook:
https://cansavvy.github.io/openpbta-notebook-concept/snv-callers/explore-tcga-pbta.nb.html

The results of this notebook were what was added to this PDF report:

TCGAvsPBTAconsensus.pdf

Reproducibility Checklist

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

Documentation Checklist

This analysis is not permanent and has not been added to the main READMEs

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

[jashapiro]

This looks pretty good, and I only have a few places where I think some DRYing would be useful.

My biggest comment is that the graphs here don't look bad? The TCGA data does seem to have higher TMB? Or am I looking at something wrong? If not, what is the difference here from previous analyses?

I am specifically looking at https://github.com/AlexsLemonade/OpenPBTA-analysis/blob/3b3fd1c3b715da590179aa40a8c13c7316c0db18/analyses/snv-callers/lancet-wxs-tests/plots/tcga-vs-pbta-plots/tmb-cdf-Consensus.png, which looks as I would have expected, vs plots that were produced previously.

[jashapiro]

I think you can just do this always, and save an argument. Taking the substr of a 12 character barcode will just return the barcode.

[jashapiro]

You have taken suggestions to my satisfaction, especially given the fact that this analysis is out of date and has been replaced by more current analyses. Given all that, I am happy to approve this PR in its current form. 🤓