Hi all,
I am trying to import a VCF with v4.5 spec, and getting the following error
Hail version: 0.2.137-733ac4ccd943
Error summary: InvalidHeader: Your input file has a malformed header: VCFv4.5 is not a supported version
I can see in the documentation that Hail is designed to be maximally compatible with files in the VCF 4.2 spec, but is there a way to make it work with later specs?
Any help here would be much appreciated.
Best,
Elliott
The issue here is hail does not implement it’s own VCF header parser. We rely on htsjdk to parse headers for us, and unfortunately, htsjdk does not yet support VCF 4.5 headers.
You may have luck simply changing the version string to 4.2 or 4.3, as hail’s VCF parsing is “loose” so to speak.
Many thanks for the quick reply. Your suggestion works with a few additional tweaks, I also had to change the number format for a few definitions, for example ##FORMAT=<ID=LF1R2,Number=LR, to ##FORMAT=<ID=LF1R2,Number=.
The vcf I am working with is from the latest Illumina Dragen IGG pipeline, it is a concern that Dragen is outputting VCFs that are not directly compatible with downsteam tools. There are many other issues with Dragen IGG VCFs that require work arounds. Given the increasing use of Dragen in biobank level and other projects, I wonder if we should establish a community led best practice resource for working with Dragen VCFs and Hail?