gnomAD: query utilitites

Question

[bsmith030465]

Hi,

I wanted to query the gnomAD database without having to download and configure the entire data. Is there a way gnomAD provides an API for programmatic access? For example, for a given variant id, I want to retrieve the Total 'Allele Count':

INPUT: '1-978628-C-T' (or 'rs116586548')
OUTPUT: 180

i.e. if we query the website (gnomAD

I have thousands of variants that I want to get the Total Allele Count for, but would prefer not to download the entire dataset.

thanks!

Answer 1

[Pierre Lindenbaum]

use an API like https://myvariant.info/

wget -q -O - "http://myvariant.info/v1/variant/rs116586548" H  "accept: application/json"  
(..)
    "an": {
      "an": 30864,
      "an_afr": 8690,
      "an_amr": 838,
      "an_asj": 302,
      "an_eas": 1622,
      "an_female": 13804,
      "an_fin": 3490,
      "an_male": 17060,
      "an_nfe": 14940,
      "an_oth": 982,
      "an_popmax": 8690,
      "an_raw": 30990
    },

(...)

Answer 2

[r-ninja]

Hi everyone, after trying many of the things suggested here, I realized none of them did what I needed. So I teamed up with a web developer and we wrote a node.js tool to do the work. It's working pretty well so far!

Github repo: https://github.com/crankysparrow/computationalbio

Blogpost: https://bbitarello.github.io/post/gnomad-api-query/#a-brief-tutorial

Basically you wanna run:

node scripts/readIDs.js --input input/rsIDs_whi.txt --output input/rsIDs_whi.json

This converts a text file with one SNP ID per line into a .json file to be used as input. Then run:

node scripts/gnomad_query.js --input input/rsIDs_jhs.json  --pop 'NFE' --db 'GNOMAD'

This actually runs http://bioinfo.hpc.cam.ac.uk/cellbase/webservices/ for you and produces an output in .json with all the queries.

Answer 3

[manuel.belmadani]

I don't think so, at least not officially. You might be able to find a third party web application that has gnomAD data.

Bravo is another variant database that seems to use part of the ExAC browser codebase: https://bravo.sph.umich.edu/freeze5/hg38/ and they provide an API. The data might not be as rich as gnomAD but it might be enough depending on what you need.

Varsome is a commerical variant database that might work here too: https://varsome.com/information/varsome-api-info/ which is free with some restrictions but also provides some premium models. They provide a sample client API as well: https://github.com/saphetor/varsome-api-client-python

There's some other solutions: https://bioinformatics.stackexchange.com/questions/933/is-there-public-restful-api-for-gnomad I couldn't get the tabix one to work though (the URL is different now, but still didn't work in my case.)

[Pierre Lindenbaum]

tabix doesn't work anymore because the server for gnomad doesn't support "range-query" anymore.

Answer 4

[Garan]

The Cellbase Swagger restful API might have what you're after, although I'm not sure how many requests it will take if you're looking to bulk query.

http://bioinfo.hpc.cam.ac.uk/cellbase/webservices/

http://bioinfo.hpc.cam.ac.uk/cellbase/webservices/rest/v4/hsapiens/feature/variation/rs116586548/info?limit=-1&skip=-1&skipCount=false&count=false&Output%20format=json

populationFrequencies   
0   
study   "GNOMAD_EXOMES"
population  "ALL"
refAllele   "C"
altAllele   "T"
refAlleleFreq   0.99945223
altAlleleFreq   0.00054777955
refHomGenotypeFreq  0.99890447
hetGenotypeFreq 0.0010955591
altHomGenotypeFreq  0
1   
study   "GNOMAD_EXOMES"
population  "OTH"
refAllele   "C"
altAllele   "T"
refAlleleFreq   0.9998051
altAlleleFreq   0.00019493178
refHomGenotypeFreq  0.9996101
hetGenotypeFreq 0.00038986356
altHomGenotypeFreq  0