You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
We used gatk HaplotypeCaller to find SNPs in our samples, following this command line: gatk HaplotypeCaller --java-options "-Xmx20g -XX:ParallelGCThreads=10" -R ref_genome.fa -I sample.bam -ERC GVCF -O output_sample.g.vcf.gz
Afterward, we also used GenomicsDBImport, GenotypeGVCFs and VariantFiltration.
However, we noticed quickly some genotypes had a GQ tag equal to 0 despite a very good DP. Looking at the PL, it seemed like gatk could not decide between a homozygous and a heterozygous site, but still called a homozygous site and displayed all reads mapping to the reference allele. Looking at the read mapping in IGV, we noticed that those sites were in fact heterozygous, and therefore miscalled.
Here is an example (we could provide more if needed) :
Site :
Here, genotype of the individual ZCh210C is said to be homozygous (GT 0/0) and have all reads mapping the reference allele (AD 35:0). The GQ is 0, and PL is 0 for both homozygous and heterozygous genotypes.
The reads in IGV for this individual map as follows:
Which clearly shows reads mapping on the alternative allele.
Why is gatk having this behavior?
Thank you for your anwsers,
Best,
Samuel
The text was updated successfully, but these errors were encountered:
Problem report
Affected tool(s) or class(es)
HaplotypeCaller
Affected version : 4.1
Description
We used gatk HaplotypeCaller to find SNPs in our samples, following this command line:
gatk HaplotypeCaller --java-options "-Xmx20g -XX:ParallelGCThreads=10" -R ref_genome.fa -I sample.bam -ERC GVCF -O output_sample.g.vcf.gz
Afterward, we also used GenomicsDBImport, GenotypeGVCFs and VariantFiltration.
However, we noticed quickly some genotypes had a GQ tag equal to 0 despite a very good DP. Looking at the PL, it seemed like gatk could not decide between a homozygous and a heterozygous site, but still called a homozygous site and displayed all reads mapping to the reference allele. Looking at the read mapping in IGV, we noticed that those sites were in fact heterozygous, and therefore miscalled.
Here is an example (we could provide more if needed) :
<style> </style>Site :
Here, genotype of the individual ZCh210C is said to be homozygous (GT 0/0) and have all reads mapping the reference allele (AD 35:0). The GQ is 0, and PL is 0 for both homozygous and heterozygous genotypes.
The reads in IGV for this individual map as follows:
Which clearly shows reads mapping on the alternative allele.
Why is gatk having this behavior?
Thank you for your anwsers,
Best,
Samuel
The text was updated successfully, but these errors were encountered: