GATK Mutect2: Would a matched normal sample be redundant with a pool of normals that included the sample?

Question

[Derrik]

I'm running Mutect2 on some paired tumor/normal samples. I've created the pool of normals and I'm about ready to run Mutect on the tumor samples, but I'm wondering if there is any added affect of including the normal sample if I'm already specifying a PoN that included the normal sample. I could be 'safe' and include the normal sample anyway, but I want to know if there's any additional effect since some of my samples do not have a matched normal, and would only have the PoN to take out germline/technical artifact. I'd like to know if the inclusion of a single matched normal file will produce any additional filtering on top of the PoN filtering. I will likely just run some tests myself and see if there's an obvious difference (and will update here for future reference), but there may be more subtle differences I'm not aware of. Thank you!

Answer 1

[Derrik]

The outputs are differing, it seems that with the paired sample, the program will call low AF variants that the unpaired run may ignore, but generally the unpaired run includes many more calls. I think for consistency I will run all my samples without their paired normal if they have it, then just subtract all variants appearing in the PoN from the outputs to focus on mutations that only appear in the tumors.