CNVkit calling strategies with matched-normal tumor

Question

[fbrundu]

I want to use cnvkit to call copy number alterations (CNAs) from Whole Exome Sequencing data (WES) of a specific tumor.

I have a set of WES tumor samples, some of which also have the matched normal sample. My question is: would be more effective, to call CNAs for each tumor, the whole set of matched normals or, if available, only the corresponding matched normal sample?

I thought that using only the single matched normal could spot only somatic variations, while using the whole set of normals would also add germline variations into account. Is that right?

Which best strategy do you suggest?

Answer 1

[Kevin Blighe]

Hey,

I presume that you are executing it with the batch command? - https://cnvkit.readthedocs.io/en/stable/pipeline.html#batch

Why not do the analysis as pooled tumour versus normal, and then separately for each tumour-normal pair? When you do the pooled analysis, you will essentially be looking for 'recurrent' CNA, whereas, with the matched tumour-normal pairs analyses, you will find sample-specific somatic CNAs, as well as the recurrent CNAs.

In Spanish and Portuguese they have a verb 'aprovechar', which means to make the most of something. Given the data that you have, analyse it in all possible ways in order to extract the maximum amount of useful information from it.

[fbrundu]

Yes, I'm using the batch command. I was thinking this was the best approach, too.

[Kevin Blighe]

Great - hope that it goes well. I assume that it's TCGA data that you've got.

[fbrundu]

Actually not, it's proprietary data. Thanks