bcftools merge, error "Could not parse the region(s)"

Question

[agathejouet]

Hi all,

I am trying to merge multiple vcf files using bcftools version 1.6. Unfortunately, I receive the following error for all of my "chromosomes":

 [E::_regions_init_string] Could not parse the region(s): chr1

ex:

[E::_regions_init_string] Could not parse the region(s): JA218_chr6:29940123..29944556_UTR-0_lenght=4434
[E::_regions_init_string] Could not parse the region(s): JA455_chr11:20047638..20052236_UTR-0_lenght=4599
[E::_regions_init_string] Could not parse the region(s): JA327_chr9:10804606..10807219_UTR-0_lenght=2614
[E::_regions_init_string] Could not parse the region(s): JA205_chr6:24945965..24949792_UTR-0_lenght=3828

I have using the following command:

bcftools merge file1.vcf.gz file2.vcf.gz file3.vcf.gz -o outfile -O v -0

My vcf files have been compressed with bgzip and indexed with tabix (also version 1.6) using:

bgzip file1.vcf; tabix -p vcf file1.vcf.gz

Not sure what is happening here. Any help would be appreciated and please, let me know if you need any additional piece of information.

Thanks very much,

Agathe

[Pierre Lindenbaum]

what is the output of

tabix --list-chroms file1.vcf.gz | head -n 50

please

[Pierre Lindenbaum]

bgzip file1.vcf; tabix -p vcf file1.vcf.gz

you'd better always force bgzip/taxix and use a logical AND

bgzip -f file1.vcf &&  tabix -f -p vcf file1.vcf.gz

[agathejouet]

The output is:

LOC_Os01g16370_chr1:9292171..9298764_UTR-0
LOC_Os01g16400_chr1:9313135..9317036_UTR-0
LOC_Os01g21240_chr1:11854526..11857948_UTR-0
LOC_Os01g25630_chr1:14525576..14530578_UTR-0
LOC_Os01g25710_chr1:14570811..14580124_UTR-0
LOC_Os01g25810_chr1:14611521..14616096_UTR-0
LOC_Os01g33684_chr1:18530856..18539774_UTR-0
LOC_Os01g35254_chr1:19515711..19521445_UTR-0
LOC_Os01g36640_chr1:20326945..20333788_UTR-0
LOC_Os01g41890_chr1:23745796..23750669_UTR-0
LOC_Os01g42330_chr1:24021015..24026491_UTR-0
LOC_Os01g52270_chr1:30047246..30048941_UTR-0
LOC_Os01g52280_chr1:30054522..30055994_UTR-0
LOC_Os01g52304_chr1:30061671..30065301_UTR-0
LOC_Os01g52320_chr1:30067560..30069086_UTR-0
LOC_Os01g57280_chr1:33097028..33103506_UTR-0
LOC_Os01g57870_chr1:33459210..33462869_UTR-0
LOC_Os01g58520_chr1:33815574..33820089_UTR-0
LOC_Os01g59340_chr1:34294608..34306006_UTR-0
LOC_Os02g04530_chr2:2011707..2018122_UTR-0
LOC_Os02g06030_chr2:3005600..3007976_UTR-0
LOC_Os02g06180_chr2:3084913..3087072_UTR-0
LOC_Os02g10900_chr2:5785295..5788769_UTR-0
LOC_Os02g16060_chr2:9136746..9140131_UTR-0
LOC_Os02g16250_chr2:9238437..9239234_UTR-0
LOC_Os02g16270_chr2:9257996..9262612_UTR-0
LOC_Os02g16330_chr2:9284786..9290815_UTR-0
LOC_Os02g17304_chr2:9922463..9928347_UTR-0
LOC_Os02g18000_chr2:10444558..10450974_UTR-0
LOC_Os02g18070_chr2:10490444..10496411_UTR-0
LOC_Os02g18140_chr2:10535633..10539697_UTR-0
LOC_Os02g18510_chr2:10776282..10780756_UTR-0
LOC_Os02g19750_chr2:11550971..11554437_UTR-0
LOC_Os02g19890_chr2:11701933..11708391_UTR-0
LOC_Os02g20420_chr2:12039379..12052089_UTR-0
LOC_Os02g26500_chr2:15556364..15560053_UTR-0
LOC_Os02g27500_chr2:16264148..16266679_UTR-0
LOC_Os02g27540_chr2:16306055..16309173_UTR-0
LOC_Os02g27680_chr2:16398585..16399403_UTR-0
LOC_Os02g41760_chr2:25097465..25099747_UTR-0
LOC_Os03g26260_chr3:15014453..15021952_UTR-0
LOC_Os03g37720_chr3:20912120..20915920_UTR-0
LOC_Os03g38250_chr3:21232110..21236446_UTR-0
LOC_Os03g48370_chr3:27529787..27536134_UTR-0
LOC_Os03g63150_chr3:35686155..35691879_UTR-0
LOC_Os04g02030_chr4:638803..642201_UTR-0

I have 759 "chromosomes" like this, in a similar format. Will also keep in mind to force overwrite and use the && (which I actually do in my rake file).

[Pierre Lindenbaum]

but is there the only whole word 'chr1' as you said in your first warning message ?

Answer 1

[Pierre Lindenbaum]

from the VCF specification : https://samtools.github.io/hts-specs/VCFv4.3.pdf

CHROM - chromosome:(...) . The colon symbol (:) must be absent from all chromosome names to avoid parsing errors when dealing with breakends.

you vcf is not valid.

[agathejouet]

Thanks very much for your help, will try to modify this!

Agathe

[Pierre Lindenbaum]

I would go for something like:

awk -F '\t' '/^#/ {print;next;} {OFS="\t";gsub(/[\:\.\-]/,"_",$1);print;}' input.vcf