Apparent loss of reads during realigment
Posted in Ask the GATK team | Last updated on

Comments (18)

Hi, I've 32 exomes in a merged BAM file, I have read groups to identify each exome, with id, sample, library and platform all set (I hope) correctly, what I can't understand why I have a discrepancy in the logs generated about the number of reads traversed by RealignmentTargetcreator Vs IndelRealigner:

From my RealignmentTargetcreator run I got:

INFO 18:24:17,286 ProgressMeter - Total runtime 5394.91 secs, 89.92 min, 1.50 hours INFO 18:24:17,286 MicroScheduler - 389,565,880 reads were filtered out during traversal out of 2,752,553,629 total (14.15%) INFO 18:24:17,287 MicroScheduler - -> 24573133 reads (0.89% of total) failing BadMateFilter INFO 18:24:17,287 MicroScheduler - -> 229871090 reads (8.35% of total) failing DuplicateReadFilter INFO 18:24:17,287 MicroScheduler - -> 135121273 reads (4.91% of total) failing MappingQualityZeroFilter INFO 18:24:17,298 MicroScheduler - -> 384 reads (0.00% of total) failing UnmappedReadFilter

Vs this from IndelRealigner

INFO 11:47:21,707 MicroScheduler - 0 reads were filtered out during traversal out of 200,100 total (0.00%)

Why have so few of the reads RealignmentTarget creator reported been traversed by IndelRealigner?

My command for IndelRealigner was:

GenomeAnalysisTK-2.4-9-g532efad/GenomeAnalysisTK.jar -T IndelRealigner --maxReadsInMemory 1000000 --maxReadsForRealignment 1000000 -known /data/GATK_bundle/hg19/Mills_and_1000G_gold_standard.indels.hg19.vcf -known /data/GATK_bundle/hg19/1000G_phase1.indels.hg19.vcf -I Merged_dedup.bam -R /data/GATK_bundle/hg19/ucsc.hg19.fasta -targetIntervals forIndelRealigner.intervals -o Merged_dedup_realigned.bam

Return to top Comment on this article in the forum