search

mcveanlab / mccortex

De novo genome assembly and multisample variant calling
https://github.com/mcveanlab/mccortex/wiki
MIT License
113 stars 25 forks source link

zero coverage genotyping #57

Closed hcdenbakker closed 7 years ago

hcdenbakker commented 7 years ago

I was looking into vcf-genotyping and introduced a zero coverage locus:

CP006053.1 4399342 . A C . PASS BUBBLE=204;K33 K33R:K33A .:. 0:0

only to find it was called as the alternative allele when genotyped:

CP006053.1 4399342 . A C . PASS BUBBLE=204;K33 GT:K33R:K33A:GQ ./.:.:.:. 1/1:0:0:0

I used this command:

./mccortex63 vcfgeno --out out.vcf.gz --kcov 30 --ploidy 2 myvcf.cov.vcf

iqbal-lab commented 7 years ago

Hey Henk

  1. in your first line you seem to have two samples, one with .:. and one with 0:0 - what does this mean? What is .:. in this line, I thought it was coverage?

  2. The entry which genotypes the sample as 1/1 also has a zero for GQ - so there is no confidence in that genotype.

hcdenbakker commented 7 years ago

re 1: .:. is a remnant of a dummy sample (needed to introduce this to get a vcf with all variable sites for an entire population), the last column represents a real sample/coverage

re 2: Overlooked the zero GQ... thanks for pointing that out. I let the 1/1 genotype call distract me.

On Thu, Apr 20, 2017 at 6:52 PM, Zamin Iqbal notifications@github.com wrote:

Hey Henk

1.

in your first line you seem to have two samples, one with .:. and one with 0:0 - what does this mean? What is .:. in this line, I thought it was coverage? 2.

The entry which genotypes the sample as 1/1 also has a zero for GQ - so there is no confidence in that genotype.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/mcveanlab/mccortex/issues/57#issuecomment-295952723, or mute the thread https://github.com/notifications/unsubscribe-auth/AND_dFk_F_PTHci5LchZb4ENp7pg2hafks5rx-GSgaJpZM4NDgOG .