This is a test version of Biostars. For the public version, visit https://www.biostars.org.
GATK3 error while performing local realignment for exome

How to fix the following error?

ERROR MESSAGE: Fasta index file /data/ngs/references/hg38gatkbundle/Homo_sapiens_assembly38.known_indels.vcf.fai for reference /data/ngs/references/hg38gatkbundle/Homo_sapiens_assembly38.known_indels.vcf does not exist

My command is:

java -jar path/to/gatk3.8.jar -T RealignerTargetCreator -R /path/to/hg38gatkbundle/Homo_sapiens_assembly38.known_indels.vcf -known /path/to/hg38gatkbundle/Mills_and_1000G_gold_standard.indels.hg38.vcf -I HG100.bam -L path/to/hg38gatkbundle/exons.hg38.bed -o HG100.IndelRealigner.intervals

NOTE: I am using hg38gatkbundle as my reference genome. hg38gatkbundle directory does not have any Homo_sapiens_assembly38.known_indels.vcf.fai for Homo_sapiens_assembly38.known_indels.vcf file

Also, exons.hg38.bed is not there in hg38gatkbundle directory

next-gen sequencing software error
• 1,910 views • link
updated by Devon Ryan Devon Ryan • written by shuksi1984 • shuksi1984

1 answer

The -R option takes a fasta file holding the reference sequence, not a VCF. In fact the only VCF that would normally be used here is the one you mentioned with --known.

As an aside, there's little purpose to realigning around indels unless you're using the old UnifiedGenotyper.

• 1 views • link
written by Devon Ryan Devon Ryan

Thanks, I changed the command accordingly and it worked. My command is :

java -jar /path/to/gatk3.8.jar -T RealignerTargetCreator -R /path/to/Homo_sapiens_assembly38.fasta -known/path/to/Homo_sapiens_assembly38.known_indels.vcf -known /path/to/Mills_and_1000G_gold_standard.indels.hg38.vcf -I HG100.bam -o HG100.IndelRealigner.intervals

But, still I didnt get the mystery of exons.hg38.bed. In the above command I removed it and it worked.

• 1 views • link
written by shuksi1984 • shuksi1984

The -L option is just restricting the program to operate over a given set of regions, it's not strictly needed though it might speed things up a bit.

• 1 views • link
written by Devon Ryan Devon Ryan

there's little purpose to realigning around indels unless you're using the old UnifiedGenotyper.

It could still be useful for non-GATK variant callers.

• 1 views • link
written by igor igor

Log in to answer this question.

More posts like this