This is a test version of Biostars. For the public version, visit https://www.biostars.org.
Somatic variant calling mutect2 ctDNA liquid biopsy

Hi!

I am trying to call ctDNA somatic variants in a sample that comes from liquid biopsy of a cancer patient. I also have a healthy control that comes from an in-house plasma sample.

I want to apply mutect2 but I have never used it before. I see I need to create a Panel of Normals to do the calling but I have some questions:

  • Does the patient liquid biopsy sample count as a normal or a tumor sample? The tumoral DNA is present but at very little concentrations.

  • Mutect2 requires at least a normal sample to analyze the tumor sample. Can I use the in-house plasma sample as the "normal sample" even though they do not come from the same individual? Shall I go for the "tumor-only" mode to generate the PoN even though its use is not encouraged?

Basically how can I generate a PoN that is valid and allows me to launch mutect correctly taking into account I have these two kinds of samples.

Thank you so much!!!!

ctdna gatk tumor mutect2 snp
• 1,661 views • link
written by bt_cepo • bt_cepo

1 answer

  1. Your purpose is to identify variants from liquid biopsy sample. Therefore you need to set this group as "tumor" group, no matter how much "tumor" proportion it contains.

  2. PoN is necessary in your case. Your normal variants in the "tumor" sample could be overwhelming. You need to filter them, right? The best practice is to use a paired normal sample. But since you don't have it, other PON data will be feasible. You can still get your tumor variants but the result could be contaminated by germline mutations or other normal variants that did not present in the public PON dataset. By establishing a PON database, please check the tutorial section provided by GATK team.

• 0 views • link
written by tomas4482 • tomas4482

Hi Tomas, thank you so much for your rapid response!

I understand from your answer that I cannot use the healthy control as the paired normal since it does not belong to the same individual.

Also, I understand that I can use a public dataset as PON. I checked the GATK resources but could only find public datasets for WES and WGS data, can I use these even though I am analyzing targeted sequencing data?

Sorry if these are dumb questions, I am only starting in bioinformatics and want to learn more! :)

• 0 views • link
written by bt_cepo • bt_cepo

Log in to answer this question.

More posts like this