[ensembl-dev] How duplicated variants in gnomAD v2 are handled in VEP when remapped to GRCh38?

[Wallace Ko]

Hello Ensembl Team,

Some variants in GRCh37 can be remapped to the same location in GRCh38.
E.g. 1-13497572-G-A (
https://gnomad.broadinstitute.org/variant/1-13497572-G-A?dataset=gnomad_r2_1)
and 1-13718406-G-A (
https://gnomad.broadinstitute.org/variant/1-13718406-G-A?dataset=gnomad_r2_1)
in GRCh37 are both remapped to 1-13391946-G-A in GRCh38 according to gnomAD
webpages.

>From the gnomAD v2 VCF in Ensembl FTP, both records are observed:

$ bcftools view -Ou
http://ftp.ensembl.org/pub/data_files/homo_sapiens/GRCh38/variation_genotype/gnomad/r2.1.1/exomes/gnomad.exomes.r2.1.1.sites.1.liftover_grch38_no_VEP.vcf.gz
-i 'ALT="A"' chr1:13391946-13391946 | bcftools query -f '%CHROM
%POS %ID %REF %ALT %AC %AN %AF\n'
chr1 13391946 rs199881782 G A 6177 85026 0.0726484
chr1 13391946 rs200047809 G A 72 211396 0.000340593

>From VEP 111, I obtained gnomADe_AF=0.0003406 for this variant.

I wonder how VEP determines to report the AF of rs200047809, instead
of rs199881782 or arithmetic mean of both.

Regards,
Wallace Ko
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.ensembl.org/pipermail/dev_ensembl.org/attachments/20240227/82ff1d37/attachment.html>