split-vep could not recognize M-CAP

[zofieLin]

Hi,

I am using bcftools +split-vep to extract the M-CAP score from the vcf annotated by VEP, but it could not recognize and produced error below:

I checked with bcftools +split-vep *.vep.vcf.gz -l , "M-CAP_score" was actually present in the CSQ field.

I also tired with some other typos, but all didn't work. Are there any solutions to solve this problem?

Thanks!

[pd3]

This is caused by the dash which is not considered a valid tag name character by VCF specification. This is a problem because +split-vep treats internally VEP fields as VCF tags and this breaks the query format expression.

A quick workaround is to edit the header and replace with underscores, e.g. M_CAP_pred. See the reheader command.

The program could be also made smarter and replace such characters automatically. Could you please show the VEP line from the header and a single data line for debugging?

[zofieLin]

Thanks for the reply.

I tried to reheader and replace it as MCAP and that works well, just need more steps to edit the header and then reheader.

Attched with some lines, see if it could help you fix this problem, M-CAP is in the last second and last third fields.
Thanks.
test.txt

[pd3]

This is now addressed, offending characters are replaced with underscores. Please try it out

[zofieLin]

Thank you so much, it worked well, with "%M_CAP_pred".

[hanguojun007]

Hi, I try reheader, but it just replace the header, not the INFO content.
when I query the new header, i can't get the value. because the header of INFO does not change.

[pd3]

@hanguojun007 This would have to be done before the plugin is run.

For example the VEP definition

##INFO=<ID=CSQ,...,Description="Format: Allele|Consequence|..|M-CAP_pred|...">

would be changed to

##INFO=<ID=CSQ,...,Description="Format: Allele|Consequence|..|M_CAP_pred|...">

with reheader, then the plugin +split-vep would work.

With the latest version this is not needed anymore.