Hi,
Sorry if this is a re-post but I have read through the last post on this issue: http://gatkforums.broadinstitute.org/gatk/discussion/4343/missing-positions-in-the-gvcf-file
I'm having the same issue except with GATK 3.6.0.
I ran haplotype caller on ~2,500 samples using the following cmd:
module load gatk/3.6.0 && module load java/1.8.0_91 && java -jar -Djava.io.tmpdir=/hpf/largeprojects/pray/llau/tmp/ -Xmx24G $GATK -T HaplotypeCaller -R /hpf/largeprojects/pray/llau/internal_databases/gatk_bundle/2.8_b37/human_g1k_v37_decoy.fasta -I .recalibrated.bam --genotyping_mode DISCOVERY -stand_emit_conf 10 -stand_call_conf 30 -rf BadCigar --dontUseSoftClippedBases --min_base_quality_score 20 --emitRefConfidence GVCF -o .raw_variants.gvcf
I then ran genotypeGVCFs using the following cmd:
module load gatk/3.6.0 && module load java/1.8.0_91 && java -jar -Djava.io.tmpdir=/hpf/largeprojects/pray/llau/tmp/ -Xmx24G $GATK -T GenotypeGVCFs --disable_auto_index_creation_and_locking_when_reading_rods --max_alternate_alleles 500 -R /hpf/largeprojects/pray/llau/internal_databases/gatk_bundle/2.8_b37/human_g1k_v37_decoy.fasta -L 20:62025520-63025520 -allSites --variant .g.vcf --variant .g.vcf..... --variant .vcf -o .vcf
If you view the attached screenshot of the .vcf file you can see I'm missing a couple of random positions (chr20: 62025529). I thought -allSites should report all the position? Am I screwing something up?