RealignerTargetCreator (GATK) contig issue

Question

[win]

So i downloaded a .bam from phase 3 1000 genomes and i want to run the RealignerTargetCreator on it using GATK 3.8.

The following command works fine:

sudo java -jar algorithms/gatk/gatk3.8.jar -T RealignerTargetCreator -R references/hg38.fasta -I data/HG100.reordered.bam -o data/HG100.realignertargetcreator.intervals

But when this command is run with -know i get contigs mismatch error between my bam and the vcf file. The know indel VCF used is located here:

https://storage.googleapis.com/genomics-public-data/resources/broad/hg38/v0/Homo_sapiens_assembly38.known_indels.vcf.gz

Step prior I have applied the picard ReorderSam using hg38.

Am i using the incorrect indels file?

Thanks in advance.

Answer 1

[Pierre Lindenbaum]

check the names, the length and the order of the contigs in

$ wget -q -O - "https://storage.googleapis.com/genomics-public-data/resources/broad/hg38/v0/Homo_sapiens_assembly38.known_indels.vcf.gz" | gunzip -c | grep '##contig'

is the same as

$ samtools view -H data/HG100.reordered.bam | grep '^@SQ'

[win]

Thanks, in the know_indels.vcf.gz i can see a lot of HLA-A, HLA-B, HLA-C contigs, where is not in the reordered BAM. Can this be an issue?

[Pierre Lindenbaum]

i can see a lot of HLA-A, HLA-B, HLA-C contigs, where is not in the reordered BAM. Can this be an issue?

yes

[win]

........and how do take care of this issue?

[Pierre Lindenbaum]

change the sequence dictionary in the bam (insert the missing lines using samtools reheader ) or remove the ##contig lines and the variants in the vcf.

[win]

OK, will try. I have never done any of this before.

[GenoMax]

Consider @Devon's instructions as you try samtools reheader: A: Problems while reheadering BAM with Samtools 1.3.1