Somatic copy number variants (CNVs) and Panel of Normals creation
Hello GATK help,
I am currently trying to understand the workflow for Somatic copy number variants (CNVs).
And I am interested to know if I only have one sample to analyze i,e 1 pair of normal & tumor exome data, can I still use to create panel of Normals or run the somatic CNV workflow to derive CNVs from exome data? Please let me know. As documentation says , I need 40 samples first to create panel of Normals then run the paired sample for copy number detection.
Please suggest.
Thanks,
Indrani
-
Using only a single sample as PON will probably won't yield any good results using GATK Somatic Copy Number variation workflow. You might want to try other tools such as Control-Freec or varscan which allow normal tumor matched CNV analysis.
-
Thank you for your earlier reply.
Is it possible to create PON with 10 -15 normal samples as the documentation says it needs 40 samples, please let me know.
Thanks
-
This would be a matter of trial and error. It will definitely show something but may contain high numbers of false positive or false negative results compared to a PON with many more samples.
-
Thanks for your reply.
Indrani
-
Hello,
Thank you for your earlier replies. I talked to my PI and we are wondering if there is any pre-built PONs for us to use if we don't want to use are own samples for PON creation. Please let me know if there is and how to use pre-built PON if I am planning to run Somatic CNVs workplace in Terra cloud platform.
Thanks,
Indrani
-
Hi Indrani Datta,
Unfortunately, we do not offer prebuilt PoNs, since the primary goal of building the PoN using similarly sequenced control samples is to learn patterns of sequencing bias and noise that are specific to your case samples.
If you do not have such controls available, you might try building a PoN from publicly available samples that you know to be similarly sequenced. However, it is difficult to guarantee performance in this scenario.
Please sign in to leave a comment.
6 comments