We've moved!
This site is now read-only. You can find our new documentation site and support forum for posting questions here.
Be sure to read our welcome blog!

Using gatk on Non-human re-seq with no dbSNP & dbIndel

HubertHubert South KoreaMember

Dear the GATK team,

I'm very appreciate the gatk team.
The gatk has been very useful analysis tool for my study on human-reseq data.

I've tried using gatk on non-human re-seq data with no dbSNP & dbIndel in these days.
In that case, I found a strange result of variant calling with low QUAL value from 10 to 60. I picked up any variant with low QUAL, and looked up it in a BAM file. But that was not real variant, i think it is because there was no variant recalibration step. So i try to filter variant calling result with some value, and here is my questions.

1) As i know it, 'QUAL' means a phred-scaled quality score assigned by the variant caller. So i tried to use it to filter the variant calling result. Is there any specific filtering threshold of QUAL for non-human re-seq data?

2) By one of the post in the gatk forum, one of the GATK Dev. member said that they recommend looking at QD, not QUAL.
( http://gatkforums.broadinstitute.org/discussion/6051/how-to-interpret-a-very-broad-distribution-of-qual )
In my case, can i use QD for filtering non-human re-seq data?
And would you recommend any value and specific threshold for filtering non-human re-seq data?

Thank you in advance.

Yours respectfully,


Best Answer


Sign In or Register to comment.