gnomad v4 in VEP about ensembl-vep HOT 14 OPEN

In the meantime, you can download the gnomAD4 VCFs and use --custom

gnomADv4.0 has both Exome and Genomes, and AFAIK there is no combined score available, so you need to sum the counts and re-calculate the AF if you want them merged

I have some scripts to do this (and cut down the INFO fields to the dozen or so I want) if you want to take a look:

https://github.com/SACGF/variantgrid/blob/master/annotation/annotation_data/generate_annotation/gnomad4.0_download.sh
https://github.com/SACGF/variantgrid/blob/master/annotation/annotation_data/generate_annotation/gnomad_data.py

from ensembl-vep.

Hi @davmlaw
thank you for this help. I appreciate it.
I tried to run the gnomad_data.py script, I already downloaded the gnomad4 data with the gnomad4.0_download.sh script; but I get syntax error on line 131 in the gnomad_data.py script

"with (open(chrom_script, "w") as cs): the arrow is under the w

SyntaxError: invalid syntax

I'll appreciate your help

from ensembl-vep.

Hi @ntm (and others interested).

We are currently not certain in which version of Ensembl we will update to the latest gnomAD, but it's likely to be 113. We tend not to rush to incorporate the initial major release version (4.0 in this case) because gnomAD typically provide an updated minor version not that long after each major release.

If you want to start using the data before we incorporate it to the cache, you can, as described in this thread, use the --custom option to use the latest data (with a bit of pre-parsing to get it in the right shape).

from ensembl-vep.

Thanks for the info @jamie-m-a , very helpful for planning ahead and deciding our course of action.
While we're talking variant frequencies, any progress on the integration of ALFA, maybe via a plugin as mentioned by @helensch ?
Thanks!

from ensembl-vep.

hi @ntm, were you able to run @davmlaw scripts? I am having issues running them. I'll appreciate you help. Please see my comment above. Thank you.

from ensembl-vep.

Hi @trust-odia - I've removed the outer brackets, which may fix the script (hard to know as it works on my machine)

You may also want to just download them from here:

https://variantgrid.com/download/annotation/VEP/annotation_data/GRCh38/gnomad4.0_GRCh38_combined_af.vcf.bgz
https://variantgrid.com/download/annotation/VEP/annotation_data/GRCh38/gnomad4.0_GRCh38_combined_af.vcf.bgz.tbi

from ensembl-vep.

You don't need to process the data if you just download the already processed/combined VCF in the comment above yours.

VEP should randomly seek inside the VCF so I don't think it should matter much how big the VCF is

If you want to make separate combined VCF files from exomes/genomes, download the individual per-chromosome files from the gnomAD site, then concatenate them, that shouldn't take much processing power either (leaving an old laptop on overnight)

from ensembl-vep.

from ensembl-vep.

from ensembl-vep.

@davmlaw

Hi Dave,

I use the combined gnomad4 data that you sent me (Thank you for this), to run VEP, just only for the AF freq, but the gnomADg_AF column does not have new values.

Please see my command:

vep -i /project/dshared/projects/VEP/chr20_NF.vcf.split.gz --tab -everything --buffer_size 1000 --offline --fork 4 --dir_plugins /project/shared/projects/PMBB_VEP_Annotations --dir_cache /home/todia --custom file=/project/shared/projects/gnomad4/gnomad4.0_GRCh38_combined_af.vcf.bgz,short_name=gnomad4,format=vcf -o /project/shared/projects/VEP/vep_out/testAnnot.chr20.txt

I will appreciate your response.

Thank you

from ensembl-vep.

@trust-odia I think you need to add fields=AFas an argument to custom - then it will add a column called gnomad4_AF (ie custom short_name + field)

See VEP Custom documentation

By default it just uses the VCF ID field, and calls the column the custom short name (ie gnomad4 in your case)

from ensembl-vep.

@davmlaw Thank you.

from ensembl-vep.

from ensembl-vep.

Hi, it's hard to know without looking at your files, but it could be due to chromosome mappings. I use the full contig names in the VCF, you may need to use --synonyms VEP Option

from ensembl-vep.