Query regarding somatic variant calling

Question

[iti.gupta]

Hello, I had a query regarding variant calling. I have three tissues from the same patient (Tumor, Non-Tumor, and Blood) I wanted to identify the somatic variants and distinguish them from the germ-line variants. How should i go about with the comparison? 1) Call the variants in all the 3 tissues separately using Haplotype caller? 2) Call Tumor and Non-Tumor using Mutect2 and Blood using Haplotype caller? 3) Make 3 ways comparison using Mutect2 with each of the tissues? Any document/post discussing this would also be helpful!! Thanks a lot in advance :)

Answer 1

[bari.ballew]

What exactly is "Non-Tumor"? To call somatic variants in your tumor sample, you should use a somatic caller (e.g. MuTect2) with the blood sample as the normal. If Non-Tumor is a nearby normal, that is, a biopsy of normal tissue taken adjacent or very close to the site of the tumor, I would not recommend using it as a normal sample. Nearby normal samples are very often "contaminated" with tumor cells that may contain tumor-specific variants.

If you also want to look at the germline variants in your subject, you can use a germline caller (e.g. HaplotyeCaller) with the blood sample.

[iti.gupta]

Yes, correctly mentioned the non-tumor is the nearby normal tissue The disease in my case is the skin de-pigmentation disorder in which loss of melanocytes occur in a patchy fashion. In such a case i doubt that contamination point is something to consider I am still not very clear can you please elaborate given the present disease scenario?