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!

Genotype caller discrepancies

henrytohenryto Member
edited August 2012 in Ask the GATK team

I have found some discrepancies in the genotype call via GATK and the read fields and would like some help to explain what these mean. This analysis was done using GATK v1.2-12-g15a410b.

a) I have found a genotype call which is not 0/0, 0/1 or 1/1 - what does this mean?
2/2:17,14,1:140:36.84:2375,2267,3131,196,37,0 (Note there is no other variant called at this position)

b) I found a variant with an extra number in the alt read field, and wonder which number is the true alt read?
0/1:147,102,1:250:99:2604,0,4675,3010,4919,7929

c) GATK has called a heterozygous call at this position in my data, but when i viewed the bam file on IGV, I could not find the same variant call.
Chr14:23548787, 0/1:202,0:202:99:2972,0,6925

d) GATK has appropriately called a het call for this variant, but the alt reads are listed as 0.
0/1:121,0:119:29.01:29,0,1997

Thanks!

Answers

  • ebanksebanks Broad InstituteMember, Broadie, Dev ✭✭✭✭

    Hi there,

    First off, you are using an extremely old version of the GATK and you should definitely update to the latest version. Secondly, all of your questions can be answered by looking through the VCF specification and/or the online documentation on this forum.

  • henrytohenryto Member

    Thanks for that. I have had a scan though the online documentation prior, but could not find anything that readily came up. If you have any search terms to suggest, that would be helpful. I will be looking through the VCF specification page in more detail.

  • Hi ebanks!

    Thanks very much for your kinds of help! I also have same question with "henryto" (********GATK has appropriately called a het call for this variant, but the alt reads are listed as 0. 0/1:121,0:119:29.01:29,0,1997********), could you please explain or show us exactly where we can read that answer? I could not find that! Thanks!

  • ebanksebanks Broad InstituteMember, Broadie, Dev ✭✭✭✭

    My answer for you is the same as the one I gave to henryto.

Sign In or Register to comment.