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!

Three allele calls for a diploid organism

jasongallant1jasongallant1 East Lansing, MIMember

Hello Everyone,

Not sure what to make of this one-- I have a diploid fish that I'm examining an interesting variant for. I generated a "bamout" file using Haplotype caller for one particular individual to examine the raw reads.

Strangely, there are at least three distinct variants present in the reads for this fish! There is either an A/G or deletion at this site. This leads haplotype caller to produces a VCF that calls an INDEL on top of the SNV (A/G), which is what motivated me to initially inspect this file. The flanking SNVs (and the vast majority of sites) are biallelic), and I have no reason to suspect my fish is triploid. Is this something artefactual produced by the strange structural variation in this location? Any parameter tweaks that folks can suggest that make this start to make sense?

In attatchment are the Artificial Haplotype read groups produced by HaplotypeCaller, and a second attatchment shows the actual reads supporting this mess. Any help would be greatly appreciated!!

Jason Gallant

Best Answer


  • SheilaSheila Broad InstituteMember, Broadie admin

    Hi Jason,

    Can you please post some zoomed out IGV screen shots of both the original BAM file and bamout file (including ~300 bases before and after the site)? Please also post the VCF record for the site. Are you using the latest version of GATK? What are the different colored reads in the second screenshot?


  • jasongallant1jasongallant1 East Lansing, MIMember

    Hi Sheila,

    Thanks for the rapid response-- sorry for the delay. Attatched are screen shots that you requested (about 400bp around the variant up and down stream).

    I am using the latest release of GATK (3.7).

    The colors represent the different read groups. (Pink is the "Artificial Haplotype" read group produced by the bamout command, which is now grey in this second set of screen shots).

    I've included a second set of screen shots for the original bam file (closeup and ~400bp up and down stream). It does appear that some form of Indel realignment has occured given that the A is not in the same uniform position as it is in the bamout file.

    I've also attatched the VCF record for the region.


Sign In or Register to comment.