Are Gene–Gene Synergy and GxE Interactions on the Open Targets Roadmap?

General
1

Hi Open Targets team and community,

I’m curious whether gene–gene interactions (synergistic effect) and gene–environment interactions (GxE) are currently on the Open Targets roadmap.

These interaction layers are increasingly important for understanding complex disease biology, especially when moving beyond single-variant or single-gene associations toward more systems-level insights. Incorporating such data could significantly enhance target prioritization, particularly for multifactorial diseases.

A few specific questions:

  • Are there plans to integrate datasets or models capturing gene–gene interactions (e.g., epistasis networks)?

  • Is there any ongoing work or consideration around gene–environment interaction data (e.g., lifestyle, exposure, or longitudinal cohort data)?

Would love to hear thoughts from the team or others in the community who may be exploring this direction.

Thanks!

Shicheng

2

Hi Shicheng,

Thanks for raising this. It is indeed one of the interesting frontiers in target prioritisation and our scientific leadership has been thinking about it too.

I suspect you are mostly interested in such data for common or complex diseases. For the record, we have been intentionally excluding GWAS summary statistics from the GWAS Catalog on GxG and GxE since they violate some of the statistical assumptions we rely on. The problem is quite complex, and we would be very happy to hear how people are thinking about using such data so we can incorporate new functionalities and capabilities.

More generally, we have been thinking about how multiple genes or proteins could be nominated as new potential drug targets in a way that is amenable for therapeutic intervention. The following are a subset of things that have already been released on the topic:

  • One of the hidden gems in the Platform is the ability to explore physically or functionally interacting genes and proteins through the associations page, available in the modal next to a target. This draws on databases such as IntAct, StringDB, Signor and Reactome, and is a good starting point for thinking about multiple genes in the context of a disease.
  • One example of how we have used genetics and molecular biology together to improve target prioritisation can be found in this work (Network expansion of genetic associations defines a pleiotropy map of human cell biology | Nature Genetics).
  • Another relevant data layer we have integrated is trans-pQTLs from the UK Biobank, which can help nominate genes or proteins that act in combination with the cis-regulated gene, offering a window into downstream interaction effects.
  • One area we are keen to develop further is exploiting the latest updates from EMBL-EBI’s AlphaFold Database, which has recently added structural models for protein complexes (Millions of protein complexes added to AlphaFold Database shed light on how proteins interact | EMBL-EBI). This opens exciting possibilities for understanding how proteins physically interact at a structural level.

We would love to hear more about how you and others in the community are approaching this, as it will help shape where we focus our efforts.

Best,
David