created by Geraldine_VdAuwera
on 2016-04-02
The Panel of Normals (PoN) plays two important roles in somatic variant analysis:
1. Exclude germline variant sites that are found in the normals to avoid calling them as potential somatic variants in the tumor;
2. Exclude technical artifacts that arise from particular techniques (eg sample preservation) and technologies (eg library capture, sequencing chemistry).
Given these roles, the most important selection criteria are the technical properties of how the normal data was generated. It’s very important to use normals that are as technically similar as possible to the tumor. Also, the samples should come from subjects that were young and healthy (to minimize the chance of using as normal a sample from someone who has an undiagnosed tumor).
If possible it is better to use normals generated from the same type of tissue because if the tissues were preserved differently, the artifact patterns may be different.
From dayzcool on 2017-06-19
I read from Broad’s presentation that PoN is preferable to matched normal. I wonder those samples you have used for PoN validation are sequenced technically consistently. (e.g. Tumor/Normal samples were sequenced using same sequencing machine and chemistry?)
I would like to try PoN for Mutect and CNV calling, but it is challenging to get many sequencing samples for PoN due to the curse of dimensionality. (e.g. sequencing machine, chemistry version, technician, etc.)
From Sheila on 2017-06-22
@dayzcool
Hi,
The point of the PoN is to eliminate technical bias, so yes you should try to get normal samples sequenced in the same way as your tumor samples. Have a look at [this thread](http://gatkforums.broadinstitute.org/gatk/discussion/7796/pon-in-mutect2/p1) for more information.
-Sheila