Quantcast
Channel: haplotypecaller — GATK-Forum
Viewing all articles
Browse latest Browse all 1335

same setting, different results (GATK realignment,recalibration,haptypecaller,hard filtering)

$
0
0

When I run GATK with identical settings on some amplicon sequencing data from MiSeq (150KB region), I get different numbers of variants (approximately 10% difference), even after setting -dfrac to 1. What is the cause for this variation? How to make results reproducible?

Thank you so much!

RESULT

#8.1.vcf and 8.2.vcf are the raw VCFs from two runs with filters applied
java -Xmx4g -jar gatk.jar -T CombineVariants -R human_g1k_v37.fasta -o 8merge_union.vcf -V 8.1.vcf -V 8.2.vcf
grep Intersection 8merge_union.vcf > 8merge_intersection.vcf
grep -v Intersection 8merge_union.vcf > 8merge_NOintersection.vcf
grep PASS 8merge_NOintersection.vcf > 8merge_NOintersection.PASS.vcf

wc -l *vcf
   711 8.1.vcf
   642 8.2.vcf
   308 8merge_intersection.vcf
    86 8merge_NOintersection.PASS.vcf
   462 8merge_NOintersection.vcf
   770 8merge_union.vcf

Please find 8.1.vcf and 8.2.vcf in attachment.

COMMAND:

##realign
java -Xmx4g -jar /home/user/Downloads/gatk2.8/GenomeAnalysisTK.jar -T RealignerTargetCreator -I 8.bam -R ~public/project/seqlib/g1k_v37/human_g1k_v37.fasta -known ~public/project/seqlib/gatk/Mills_and_1000G_gold_standard.indels.b37.vcf -o 8.intervals -nt 6 -L /home/user/projects/data/collaborator_enzyme_compare/collaborator_capture2.bed -dfrac 1
java -Xmx4g -jar /home/user/Downloads/gatk2.8/GenomeAnalysisTK.jar -T IndelRealigner -I 8.bam -R ~public/project/seqlib/g1k_v37/human_g1k_v37.fasta -targetIntervals 8.intervals --out 8.realign.bam -known ~public/project/seqlib/gatk/Mills_and_1000G_gold_standard.indels.b37.vcf --consensusDeterminationModel USE_READS -LOD 0.4 -L /home/user/projects/data/collaborator_enzyme_compare/collaborator_capture2.bed -dfrac 1
#
##base recal
java -Xmx4g -jar /home/user/Downloads/gatk2.8/GenomeAnalysisTK.jar -T BaseRecalibrator -I 8.realign.bam -R ~public/project/seqlib/g1k_v37/human_g1k_v37.fasta --default_platform ILLUMINA -knownSites ~public/project/seqlib/gatk/dbsnp_137.b37.vcf -knownSites ~public/project/seqlib/gatk/Mills_and_1000G_gold_standard.indels.b37.vcf -nct 6 -o 8.recal_data -L /home/user/projects/data/collaborator_enzyme_compare/collaborator_capture2.bed -dfrac 1
java -Xmx6g -jar /home/user/Downloads/gatk2.8/GenomeAnalysisTK.jar -T PrintReads -I 8.realign.bam -R ~public/project/seqlib/g1k_v37/human_g1k_v37.fasta -o 8.recal.bam -BQSR 8.recal_data -nct 1  -L /home/user/projects/data/collaborator_enzyme_compare/collaborator_capture2.bed -dfrac 1

#variant calling
java -Xmx8g -jar /home/user/Downloads/gatk2.8/GenomeAnalysisTK.jar -T HaplotypeCaller  -R ~public/project/seqlib/g1k_v37/human_g1k_v37.fasta -I 8.recal.bam -o 8.raw.vcf --dbsnp ~public/project/seqlib/gatk/dbsnp_137.b37.vcf -nct 6 -L /home/user/projects/data/collaborator_enzyme_compare/collaborator_capture2.bed -dfrac 1

#variant filtering
java -Xmx4g -jar /home/user/Downloads/gatk2.8/GenomeAnalysisTK.jar -T SelectVariants -R ~public/project/seqlib/g1k_v37/human_g1k_v37.fasta --variant 8.raw.vcf -o 8.raw.indel.vcf -selectType INDEL -L /home/user/projects/data/collaborator_enzyme_compare/collaborator_capture2.bed -dfrac 1

java -Xmx4g -jar /home/user/Downloads/gatk2.8/GenomeAnalysisTK.jar -T SelectVariants -R ~public/project/seqlib/g1k_v37/human_g1k_v37.fasta --variant 8.raw.vcf -o 8.raw.snp.vcf -selectType SNP -L /home/user/projects/data/collaborator_enzyme_compare/collaborator_capture2.bed -dfrac 1

#use hard filtering due to small input file
#SNP
java -Xmx4g -jar /home/user/Downloads/gatk2.8/GenomeAnalysisTK.jar -T VariantFiltration -R ~public/project/seqlib/g1k_v37/human_g1k_v37.fasta --variant 8.raw.snp.vcf --filterName QDFilter --filterExpression 'QD<2.0' --filterName FSFilter --filterExpression 'FS>60.0' --filterName MQFilter --filterExpression 'MQ<40.0' --filterName MaqQualRankSumFilter --filterExpression 'MappingQualityRankSum<-12.5' --filterName ReadPosFilter --filterExpression 'ReadPosRankSum<-8.0' -o 8.filter.snp.vcf -L /home/user/projects/data/collaborator_enzyme_compare/collaborator_capture2.bed -dfrac 1
#indel
java -Xmx4g -jar /home/user/Downloads/gatk2.8/GenomeAnalysisTK.jar -T VariantFiltration -R ~public/project/seqlib/g1k_v37/human_g1k_v37.fasta --variant 8.raw.indel.vcf --clusterWindowSize 10 --filterExpression 'QD<2.0' --filterName QDFilter --filterExpression 'ReadPosRankSum<-20.0' --filterName ReadPosFilter --filterExpression 'FS>200.0' --filterName FSFilter -o 8.filter.indel.vcf -L /home/user/projects/data/collaborator_enzyme_compare/collaborator_capture2.bed -dfrac 1
java -Xmx4g -jar /home/user/Downloads/gatk2.8/GenomeAnalysisTK.jar -T CombineVariants -R ~public/project/seqlib/g1k_v37/human_g1k_v37.fasta --variant 8.filter.snp.vcf --variant 8.filter.indel.vcf -o 8.filter.vcf -L /home/user/projects/data/collaborator_enzyme_compare/collaborator_capture2.bed -dfrac 1

Viewing all articles
Browse latest Browse all 1335

Trending Articles