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!

A question about using GenomeAnalysisTK to deal with ancient DNA data

I want to call variants for some ancient samples using HaplotypeCaller.
My command is as follows:
java -jar GenomeAnalysisTK.jar -R myref.fa -T HaplotypeCaller -ERC BP_RESOLUTION -I 1.bam -L site.list -o 1.genotype.g.vcf.gz
java -jar GenomeAnalysisTK.jar -R myref.fa -T HaplotypeCaller -ERC BP_RESOLUTION -I 2.bam -L site.list -o 2.genotype.g.vcf.gz
java -jar GenomeAnalysisTK.jar -R myref.fa -T HaplotypeCaller -ERC BP_RESOLUTION -I 3.bam -L site.list -o 3.genotype.g.vcf.gz
java -jar GenomeAnalysisTK.jar -R myref.fa -T CombineGVCFs --variant 1.genotype.g.vcf.gz --variant 2.genotype.g.vcf.gz --variant 3.genotype.g.vcf.gz -o merge.g.vcf.gz
java -jar GenomeAnalysisTK.jar -R myref.fa -T GenotypeGVCFs --variant merge.g.vcf.gz --includeNonVariantSites -stand_call_conf 30 -stand_emit_conf 30 -o merge.vcf.gz
In the 3.genotype.g.vcf.gz ,a site is :
1 72710 . G C, 2.99 . DP=1;ExcessHet=3.0103;MLEAC=1,0;MLEAF=0.500,0.00;RAW_MQ=3600.00 GT:AD:DP:GQ:PL:SB 1/1:0,1,0:1:3:25,3,0,25,3,25:0,0,1,0
In the merge.g.vcf.gz ,the site is :
1 72710 . G C, . . DP=1;ExcessHet=3.01;RAW_MQ=3600.00 GT:AD:DP:GQ:PL:SB ./.:0,0,0:0:0:0,0,0,0,0,0 ./.:0,0,0:0:0:0,0,0,0,0,0 ./.:0,1,0:1:3:25,3,0,25,3,25:0,0,1,0
But in the merge.vcf.gz ,the site is :
1 72710 . G C . . AC=0;AF=0.00;AN=2;DP=1;ExcessHet=3.01;MQ=60.00 GT:AD:DP:RGQ ./.:0,0:0:0 ./.:0,0:0:0 0/0:0,1:1:3
Why the site with low QUAL changes from 1/1 to 0/0 finally but doesn't delete it ?
What kind of impact is the result of this for me ?
Because my variantsite is few,every site is important.
Could you suggest how to fix this problem?


  • Geraldine_VdAuweraGeraldine_VdAuwera Cambridge, MAMember, Administrator, Broadie admin
    You simply do not have enough coverage at this site to make useable calls. Here the program is still emitting the site because of the no-calls in two samples (by default we want to see where the program was uncertain) but in any case you cannot do anything with this kind of site. You can filter it out in the next step.
Sign In or Register to comment.