VCF 4.2

[DR]

I have some VCF files in version 4.2 that were produced by samtools mpileup and bcftools view. When I try to load them I get "VCFv4.2 is not a supported version".

I convert them to 4.1 via
    cat 42.vcf | vcf-convert -v 4.1 > 41.vcf
It produces the file but outputs "Downgrading of VCF versions is experimental: expect troubles!"

When I try to load this 41 vcf, IGV says "Unable to parse header with error: Your input file has a malformed header: unexpected tag count 5 in line  , for input source".

Any suggestions on what I should do?

[Jim Robinson]

Hi,  I will investigate with the GATK team that develops the VCF reader.  Apologies for the trouble.

Jim

[Stéphane Plaisance]

I second yourhelp request, having a similar issue.

When calling vcf-convert v4.1 with a 4.2 file, the software crashes with

Downgrading of VCF versions is experimental: expect troubles!

Sanity check failed: the ref does not agree at chr1:1600091 .. [] in .fa,  [T] in .vcf

 at /usr/local/biotools/vcftools/bin/vcf-convert line 21, <STDIN> line 1775.

        main::error('Sanity check failed: the ref does not agree at chr1:1600091 ....') called at /usr/local/biotools/vcftools/bin/vcf-convert line 155

        main::convert_file('HASH(0x220bd48)') called at /usr/local/biotools/vcftools/bin/vcf-convert line 12

although the fasta file was freshly generated from the very bowtie2 index used for mapping the reads in the 1st place!!!

What I am trying:

I called with bcftools1.0 htslib that generates v4.2 and with varscan2 generating v4.1

I can force ignore the version difference but have no idea what this could bring as bias.

* is it a problem to vcf-isec 4.1 with 4.2 with -f (at the level of th eresult of course)

* is there any advance in fixing vcf-convert to work down to 4.1 (or any other manual edits to obtain this)?

Thanks for your help

Any suggestions from the developers on what I should do would be really welcome

[Jim Robinson]

Hi all,

Please try the development/nightly build, it uses an updated "htsjdk" library.  If you get the same (or any) errors with that I'll take it up with the htsjdk team.

Thanks

Jim

--

---
You received this message because you are subscribed to the Google Groups "igv-help" group.
To unsubscribe from this group and stop receiving emails from it, send an email to igv-help+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/igv-help/6a3abc6e-b399-4840-a9fc-256b1873f7e0%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.