How to deside gene CNVs from .cnr outputs

Question

[guisongtao0104]

Dear all,

I have run cnvkit using WGS data, and gene regions were provided as targets. And I get output.cnr like this:

chr     start   end       gene           log2         cn        depth   weight
1       44288   46137   gene027230     -0.126498       2       11.2612 0.805428
1       46137   47987   gene027230     -0.00296269     2       12.9805 0.822361
1       47987   49837   gene027230     -0.0577102      2       12.9541 0.831048
1       50876   52489   gene027231     -0.0192799      2       13.509  0.787347
1       52489   54102   gene027231     0.0389804       2       13.9628 0.788384
1       54102   55716   gene027231     -0.0985898      2       11.0235 0.722622
1       92298   93716   gene027232     -0.365783       1       11.0071 0.739006
1       93716   95134   gene027232     -0.037821       2       13.2454 0.754825

And I would like to know which gene is duplicated/deleted. How could I merge log2-values of bins that belong to the same gene? And is it reasonable that genes with log2-value > 1.0 were considered as duplications, and those with log2-value < −1.0 constituted deletions.

Thank you for your help

Songtao Gui

Answer 1

[Chip]

To merge bins you need to run cnvkit.py segment to create .cns files.

Assuming a diploid genome:

• Log2(CR) = -1 → CR = 0.5 → 1 copy (of a segment/gene).
• Log2(CR) = 0 → CR = 1 → copies
• Log2(CR) = +1 → = 2 → 4 copies
• Log2(CR) = 0.58 → = 1.5 → copies

Of course you have to consider a bit of margin around those values to call deletions or amplifications: for example, you can assume that with Log2(CR) = -1.1 your organism still have you gene copy and with Log2(CR) = 0.4 you can be confident that it has 3 gene copies.