upstream_gene_variant with distance 0

[0xaf1f]

Describe the issue

We noticed a variant, a single base insertion after the first position of a gene, getting annotated as an upstream_gene_variant with respect to that gene. The distance is then 0 since it's not actually upstream, but why isn't it instead called a frameshift mutation?

This is the problematic annotation:

A|upstream_gene_variant|MODIFIER|dxs2|Rv3379c|Transcript|CCP46200|protein_coding|||||||||||0|-1||1|insertion|ENA_GENE|

Additional information

System

• VEP version: 105.0
• VEP Cache version: 30
• Perl version: 5.32
• OS: Debian
• tabix installed ? yes

Full VEP command line

vep --force_overwrite --dir /path/to/.vep --synonyms /path/to/.vep/synonyms.txt --offline --cache --cache_version 30 --species mycobacterium_tuberculosis_h37rv --symbol --variant_class --flag_pick --vcf -i test-case.vcf -o test-case.annotated.vcf

Full error message

N/A

Data files (if applicable)

vep-distance-issue.tar.gz

[diegomscoelho]

Hi @0xaf1f,

We are currently investigating your issue. I will post an answer here shortly.

Thanks for your question.

Regards, @diegomscoelho

[nakib103]

Hello @0xaf1f,

Sorry for delay in re-visiting this issue. I cannot check the exact example you provided but can re-produce the case in human GRCh38 with input -

1   230714122   .   C   CT

results in upstream_gene_variant with DISTANCE=0

#Uploaded_variation Location    Allele  Gene    Feature Feature_type    Consequence cDNA_position   CDS_position    Protein_position    Amino_acids Codons  Existing_variation  Extra
1_230714123_-/T 1:230714122-230714123   T   ENSG00000135744 ENST00000366667 Transcript  upstream_gene_variant   -   -   -   -   -   -   IMPACT=MODIFIER;DISTANCE=0;STRAND=-1

For insertions it makes sense. As the variant position is considered to be the flanking bases where the insertion actually happens (see here) making start (or end) of the variant position same as transcript start (or end). The inserted sequence itself is outside the transcript so it is correct to say the effect is upstream.

I have added a PR to make it clear in the doc here - https://www.ensembl.org/info/docs/tools/vep/vep_formats.html#defaultout

As the issue has been stale for long time I will close it.

Best regards, Nakib

[0xaf1f]

Thanks -- in my example data, the gene for which this was annotated as distance 0 was has coordinates 3793257-3794867 on the minus strand.

Because the variant is

1       3794867 .       C       CA 

you're right that it is in fact upstream-- just before the start. Although the distance=0 is still odd. I'd have expected a distance 1 perhaps, but I suppose that's not as big a deal.