different results between jointcalling from gvcfs and bams Answered

Post author


I used different method to do jointcalling: 


gatk HaplotypeCaller --native-pair-hmm-threads 20 -ploidy 1 -I 1.bam -I 2.bam -I 3.bam -O vcf.gz


gatk HaplotypeCaller --native-pair-hmm-threads 20 -ploidy 1 -I ${bam} -R ${ref} -ERC GVCF --tmp-dir ./ -O ${sample_id}.g.vcf.gz

gatk CombineGVCFs -R ${ref} -V ${sep =' -V ' gvcfs} -O ${project_id}.raw.g.vcf.gz

gatk GenotypeGVCFs -R ${ref} -V ${project_id}.raw.g.vcf.gz -O ${project_id}.raw.vcf.gz


then processed with same FilterVariation

In g.vcf method's result, there were 2 variations that labeled "PASS" which not "PASS" in bam's result.

I don't know which result is better.

I will appreciate if you can help to answer it .




1 comment

  • Comment author
    Jason Cerrato

    Hi weikangfei,

    Thank you for your question. Our GATK team will be better suited to help you here! I recommend searching the GATK forum to see if this question has been answered before. If not, you can ask a new question which the GATK team will try to assist with as soon as they are able.

    Please be sure to read the Read Before You Post link before posting.


    Kind regards,



Please sign in to leave a comment.