Proposed Analysis: visualization of CNV and SV data with Circos plot

[jaclyn-taroni]

Related to the following: #393, #394, https://github.com/AlexsLemonade/OpenPBTA-analysis/issues/392#issuecomment-570332277, and https://github.com/AlexsLemonade/OpenPBTA-analysis/issues/27#issuecomment-516471524

What are the scientific goals of the analysis?

To generate circular plots that integrate CNV and SV data. Specifically, it might be useful for the following cases:

• Display examples of chromothripsis per https://github.com/AlexsLemonade/OpenPBTA-analysis/issues/27#issuecomment-516471524
• There are some subtype tickets that would benefit from such a visualization (e.g., "[ST-EPN-YAP1] Focal 11q most frequent CNA, with rest of genome stable" #245)
• Likely most useful to examine translocations

The implementation should focus on flexibility and extensibility because we don't know what "final" output will be required for the project.

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

Here I'm including links to code that could potentially be used to guide this analysis, rather than a specific method.

From https://github.com/AlexsLemonade/OpenPBTA-analysis/issues/392#issuecomment-570332277:

I've found this UMCCR pipeline code that uses Manta and CNVkit output to create circos plots. I bet we can adapt it (or use it to guide something similar) to use the "consensus filtered" CNVkit data mentioned here, the AnnotSV-annotated Manta file, and circlize.

Relevant functions from @gonzolgarcia's package mentioned in https://github.com/AlexsLemonade/OpenPBTA-analysis/issues/394#issuecomment-570589012:

• https://github.com/gonzolgarcia/svcnvplus/blob/ea43a5dbd7ae38f1b26520a31af1daddadbaf5dd/R/circ.chromo.plot.r
• https://github.com/gonzolgarcia/svcnvplus/blob/ea43a5dbd7ae38f1b26520a31af1daddadbaf5dd/R/circ.wg.plot.r

What input data are required for this analysis?

• SEG file, most likely the "consensus filtered" file #392 (related: #128)
• pbta-sv-manta.tsv.gz

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

~1 week is my rough estimate

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

Unknown/unassigned at this time

[cansavvy]

Since I've been working with the CNV and SV data a bit and have made circos plots before, I'm going to work on this ticket.

[cansavvy]

My general plan I'm starting on is to make a general purpose circos plotting function. What data would like to be able to be processed into a circos plot? Secondly what kinds of features will we want on these circos plots?

Data types to map:

• Chromosomal breakpoint data
• SV seg file
• CNV seg file
• SNV data???

Features to include:

• gene annotation for peaks?
• filter by subtype or histology (perhaps make subtypes each their own track)?

[cansavvy]

Some circos plotting functions and examples of how to use them have been added. What's left for this issue is to decide what we want to be plotted.

[jaclyn-taroni]

As @cansavvy stated, we have some functionality for generating Circos plots in the chromosomal-instability module. I am going to close this in favor of opening more specific issues for what, if any, specific plots should be generated.