Proposed Analysis: Chromothripsis analysis with ShatterSeek

[LauraEgolf]

What are the scientific goals of the analysis?

• Identify candidate chromothripsis regions based on SV and CNV calls
• Classify tumors as +/- chromothripsis
• Plot chromothripsis incidence by tumor subtype

What methods do you plan to use to accomplish the scientific goals?

ShatterSeek (already installed in OpenPBTA Docker). I plan to build off of the analysis started for #393. I will create a new analysis module, but I may adapt code from the existing sv-analysis module. I will most likely apply 01-process-sv-file.R, which was reviewed here, for preprocessing. Other code may also be adapted depending on its interpretability with limited documentation.

What input data are required for this analysis?

pbta-sv-manta.tsv.gz
pbta-cnv-consensus.seg.gz

How long do you expect is needed to complete the analysis? Will it be a multi-step analysis?

I aim to complete this within a week, but it may depend on how accurately chromothripsis regions can be identified based on criteria recommended by the authors. A subset of candidate regions will be manually inspected to evaluate accuracy.

Who will complete the analysis (please add a GitHub handle here if relevant)?

@LauraEgolf

What relevant scientific literature relates to this analysis?

Cortés-Ciriano, I., Lee, J.JK., Xi, R. et al. Comprehensive analysis of chromothripsis in 2,658 human cancers using whole-genome sequencing. Nat Genet 52, 331–341 (2020). https://doi.org/10.1038/s41588-019-0576-7

[LauraEgolf]

Hi all (@jharenza, @jaclyn-taroni, others) - update and question about this issue. I'm mostly finished with the analysis and I did adapt code from the existing sv-analysis module, which already underwent review. When I submit the first pull request, I was wondering if anyone has a preference whether I start a new module (I currently have it named chromothripsis) or just modify the existing module in my branch. I used 01-process-sv-file.R as-is, and I adapted some code from 02-shatterseek.R. I can modify my code to work in either module without too much trouble. I'm planning to submit one script per PR (not including the first script since it's unmodified). Thanks!

[jaclyn-taroni]

Hi @LauraEgolf, thank you for the update! My preference would probably be to modify sv-analysis because then 01-process-sv-file.R will not be duplicated in the repository. Because we're using version control, I am not particularly concerned with retaining the current state of sv-analysis in master.

If you feel like chromothripsis is a more fitting module name (and it very well may be!), the first or last PR you file could be to change the folder name. (The reason I am suggesting keeping the module name change separate from all other changes is because then the diff in GitHub will show that the changes are just files being renamed.)

I'm planning to submit one script per PR (not including the first script since it's unmodified). Thanks!

This plan sounds great! Since you weren't planning to include the unmodified script, that is another point in favor of altering sv-analysis instead of adding a new module with duplicated code.

Let me know if you have any questions!

[jharenza]

closing with PRs #1069 #1083 #1098 #1106 #1170