somatic CNV PoN recommendation?
Answered
is there a recommended source of somatic CNV PoN that works with the latest GATK?
(I know I can build my own, but just wonder if one from a larger cohort already exists)
-
Hi Qing Zhang,
There are public PoN resources available in the Resource Bundle also referenced in this article about the Panel of Normals. Does this help answer your question?
Kind regards,
Pamela
-
Hi,
I want hg38 CNV PoN, not SNV. Is there a public CNV PoN for somatic analysis?
-
Hi Qing Zhang,
All of the available public resources offered by GATK are located in the Resource Bundle. If what you are looking for is not present there, then it is not offered directly from GATK, but you may be able to find resources by searching old forum posts or from input by other users.
Kind regards,
Pamela
-
Hi, I found this file in the resource bundle https://storage.googleapis.com/gatk-test-data/cnv/somatic/wes-do-gc.pon.hdf5
-
Thank you for posting this resource Ibel Carri!
-
Hi,
Do i need to use a PON of male samples or female samples ?
And i use a PON of male samples do i need to cut in half the CNV on Y and X ?
Thank you --
-
It should not matter as the model works on any copy ratio.
-
Yes, but the ploidy calculation for male individuals is not the same. Is the CNV on the Y halved?
-
Of course. Ploidy is used as a prior for later CNV calculations to detect if there is a deletion or duplication.
-
so CNVs on the Y chromosome that are greater than 1 will be considered as DUPs and those less than 1 as DELs.
-
Exactly.
-
in the same way for the X chromosome of the male individual.
Whereas for the other chromosomes the normality threshold is 2. -
Ploidy estimation is critical for understanding where the baseline for each chromosome is therefore it would be unwise to assume that each contig has equal presence for any individual. This is also the same for germline calls where individuals with trisomy for certain chromosomes are taken into considereation for deletions and duplications within the same chromosome. Of course this estimation is not 100% perfect and may need parameter adjustment for distinguishing male and females properly under different sequencing technologies kits etc.
Please sign in to leave a comment.
13 comments