[ensembl-dev] [VEP] Bug stripping REF/ALT alleles ?

[William McLaren]

As the documentation for -minimal mentions you should use -allele_number to track alleles from input to output.

As you note for your example the ALT will be “-“ in all cases, but the start/end coordinates used to calculate the consequences will be different.

It’s a matter of opinion unfortunately whether -minimal should be the default or not; we err on the side of trying not to confuse users too much by making it not the default. It’s something that could be up for consideration for change in the future, though.

Will


On 14 July 2017 at 12:28:46, João Eiras (joao.eiras at gmail.com) wrote:

On 14 July 2017 at 12:15, William McLaren <wm2 at ebi.ac.uk> wrote:
> Hi João,
>
> VEP trims alleles if the substitution is unbalanced (typically an insertion
> or deletion where the length(REF) != length(ALT)), and by default only trims
> the first base from each.
>
> You may force VEP to fully strip each pair of alleles using --minimal [1];
> using this flag I get missense as expected.
>
> We’re currently working on expanding and clarifying the documentation for
> this, with the update available in release 90.
>
> Regards
>

Hi William.

Indeed that option fixes the annotation, but it breaks other things.
I'm parsing the JSON output and using the variant_allele property to
match an annotation with the input allele. But variant_allele is
missing in deletes so the match cannot happen. But meanwhile I managed
to get the code working with "codons".

I think the minimal option shouldn't be there as it breaks deletes
awfully. For instance, how to strip "chr1 99772782 ACTG ACT,ATG,ACG"?
The alt will always be empty.

I also think that VEP should do the stripping internally always when
producing the annotation. Whether minimal is present or not should
have no influence in the annotation.

Thank you for your time !
_______________________________________________
Dev mailing list Dev at ensembl.org
Posting guidelines and subscribe/unsubscribe info: http://lists.ensembl.org/mailman/listinfo/dev
Ensembl Blog: http://www.ensembl.info/
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.ensembl.org/pipermail/dev_ensembl.org/attachments/20170714/b40a178b/attachment.html>