Wrong predict consequence in variants using VEP Ensembl

Question

[theodorobiotec]

Hi,

I've been using Ensembl Variant Effect Predictor (VEP) to determine the effect of variants in non-model organism. Despite of being a collection of variants identified in non-model organisms it has a reference genome and a VEP database hosted in ensembl fungi portal.

So, I ran vep using a multi-sample VCF as input and got a final VCF with annotations about the consequence of each variant. When I was looking at the results I realised a variant that supposedly result in a stop codon insertion in some samples as you can check in first line bellow.

Allele      Consequence   IMPACT     SYMBOL             Gene Feature_type          Feature        BIOTYPE EXON
2055      T      stop_gained     HIGH ATEG_09961 CADATEAG00000507   Transcript CADATEAT00000507 protein_coding  8/8
2059      C missense_variant MODERATE ATEG_09961 CADATEAG00000507   Transcript CADATEAT00000507 protein_coding  8/8
     INTRON HGVSc HGVSp cDNA_position CDS_position Protein_position Amino_acids  Codons Existing_variation DISTANCE
2055                             8260         8260             2754         Q/* Cag/Tag                            
2059                             8262         8262             2754         Q/H caG/caC                            
     STRAND SYMBOL_SOURCE
2055      1         CADRE
2059      1         CADRE

But if you pay attention there's also a second variant into same codon (same Protein_position) but two bases apart from the first variant. Don't VEP should take into account the variants into same codon before output the consequence results?

Because if we consider the second's variant role and supposing that a sample is homozygote for variant allele in both variants the consequence will be the insertion of a TaC (translate to tyrosine).

I don't know if I did something wrong or miss any parameter in my analysis but it sounds a bit strange for me that VEP's algorithm doesn't handle with issues related to variants located at same codon.

Answer 1

[Emily]

The VEP does not consider variant phasing and only considers each variant individually.

Answer 2

[Christian]

We built CooVar a while back for exactly that purpose/reason (and uncovered a few discrepancies with complex variants in VEP of which I don't know if they have been resolved).

[John]

Very cool - well done for creating this! :)