Panel of tumors for CNV detection

Question

[biobiu]

Hi, I'm trying to detect somatic copy-number alterations in custom-targeted sequencing panel.

Most of the tools and methods published for this sake requires matched normal sample or panel of normals. However, we have no normal samples. We do have dozens of tumor samples sequenced with the same targets on with the same library prep methods, so, I believe we will be able to reduce the biases introduced by the target capturing and library preparation.

Did anyone tried to call somatic CNV with such "panel of tumors"? or familiar with tools that are less sensitive to this kind of setting? Best!

[biobiu]

Following further reading- I think I'll try to use CNVkit, gatk and copywriteR. Did anyone tried these tools with such "panel of tumors"? or know which should be less sensitive for this kind of input?

[pedricolino]

I am now trying to do the same, i.e. detecting CNV from panel seq. data without any kind of normals, and evaluating whether the results are useable. May I ask what you found out while trying out these tools? I am also looking at PureCN for this.

Answer 1

[lffu_0032]

In order to call CNV using catpure daga, you must need the normal or panel or normal data. the tumor data is so complex for its purity and ploidy, so the tumor data is not suit for being the control.