Tool:Open Targets Genetics: variant-centric resource for drug target identification and prioritisation

Question

[Denise CS]

Looking for a resource to help you identify causal associations between genetic variants, genes, and diseases in humans? Would you like this information to be open source, open access, and available in an easy-to-use web interface?

Open Targets has launched Open Targets Genetics, a variant-centric portal that allows you to:

• Prioritise genes using an integrated functional score
• Generate a list of genes for which a variant is functionally implicated
• Identify variants tagging a trait-associated lead variant through fine mapping or LD
• Identify traits associated with a gene or variant
• Find shared susceptibility loci
• Visualise associations between traits, variants, and genes
• Connect genes with known drugs in the Open Targets Platform

This new resource will complement the Open Targets Platform and assist with the identification and prioritisation of potential new drug targets by highlighting candidate causal variants and their implicated genes.

Check our short animation to whiz through UK Biobank data, GWAS, GTEx, FANTOM5, PCHi-C data and plenty more.

What are the entry points into Open Targets Genetics?

• Variant (either as a rsID or GRCh38 genomic coordinates e.g. 1_154453788_C_T)
• Gene (HGNC official symbol or Ensembl gene ID e.g. PCSK9)
• Trait or study, including GWAS Catalog and UK Biobank traits e.g. LDL cholesterol (Willer CJ (2013) Nat Genet)

Where does the data come from?

We integrate publicly available functional genomics datasets including eQTL, pQTL, enhancer-TSS, DHS-promoter, and promoter capture Hi-C data, as well as functional consequences predicted by Ensembl VEP and distance to TSS (transcript start site). More details can be found on our data sources documentation. The variant-phenotype associations are derived from UK Biobank disease-trait summary statistics and GWAS curated studies. We use the GWAS (lead) disease-associated variants and expand this set into causal tag variants through linkage disequilibrium (LD) expansion or fine-mapping.

Want to know more?

Head to our docs to get an overview of our approach, the technologies we use, our FAQs, and plenty more. For our release notes, check the Open Targets blog.

Have a question or want to make a comment?

Email us and we will get back as soon as we can. Sign up to our newsletter if you want to keep up-to-date with the developments and news on Open Targets Genetics.

[dagarfield]

Is there a method for accessing Open Targets Genetics programatically (e.g. API)? The API for gene-centric queries to Open Targets is great, but I'm struggling to find something similar for Open Targets Genetics from the variant perspective. Thanks!

[Denise CS]

Yes, there is the GraphQL. You can also Download the data from different locations e.g. FTP.

We are currently working on some tutorials for the GraqhQL. We have one example on querying the Open Targets Genetics API with curl and this is what is available at the moment but more will be available in our docs shortly.

If you have a specific use case, let us know and we can include that as well.

[dagarfield]

Hi again! We puzzled through it with the help of the browser at https://genetics-api.opentargets.io/ and I'm please to say that after some initially puzzling, we managed to get what we needed. Namely:

A) Look up a variant ID via the rsID { search(queryString: "rs4129267") { variants{ rsId id } } }

And B) pull out the V2G information for a given variant { genesForVariant(variantId: "1_154453788_C_T") { gene { id } overallScore qtls { typeId sourceId aggregatedScore tissues { tissue { id } beta pval } } intervals { typeId sourceId aggregatedScore } functionalPredictions { sourceId typeId aggregatedScore tissues { tissue { id } maxEffectLabel maxEffectScore } } overallScore distances { typeId sourceId aggregatedScore tissues { distance } } } }

Hopefully these will be useful to the next person in starting a graphql adventure.

[holger.klein]

Great, thanks a lot dagarfield. To make it even simpler for the next person to look into this, here's a chunk of R code to do the same (part A only, but this should get you started).

library(jsonlite)
library(httr)
req <- httr::POST("https://genetics-api.opentargets.io/graphql",
  httr::add_headers(
    "Content-Type" = "application/json"
  ),
  body = "{\"query\":\"{search(queryString: \\\"rs4899553\\\") {variants{rsId id }}}\"}"
);
res<-jsonlite::fromJSON(httr::content(req, as = "text"))

[Denise CS]

We've made a start on a GraphQL API tutorials page in the Open Targets Genetics docs.

Let us know if it's useful and if not, how to improve these. Email us at maito:[e-mail]