Our Science
Biology first.AI at scale.
We start with human disease biology, use AI to see patterns no team could find by hand, and let rigorous experiments decide what is true.
- Human-genetics anchoredTargets with causal human evidence are more likely to succeed in the clinic.
- Experimentally groundedOur own laboratory validates every key prediction.
- Modality-agnosticWe choose the modality that fits the biology, not the other way round.
- Scientists across AI, chemistry and biology
- 35
- Active discovery programmesPlaceholder
- 8
- Experimental success rate*
- 76%
* Share of AI-prioritised hypotheses confirmed in first-round wet-lab validation (owner-reported).
The problem
Most drugs fail for reasons that were knowable
Roughly nine in ten drug candidates that enter clinical trials never reach patients. The most common reasons are lack of efficacy and unacceptable safety — often traceable to a target that was not causally linked to the disease, or a molecule whose liabilities were not visible early.
Our science is built to surface that evidence before the expensive steps: causal human genetics for targets, multi-parameter design for molecules, and early experimental tests of both.
Research themes
What our scientists work on
Structure-aware generative chemistry
Diffusion and graph models conditioned on binding-site geometry.
Active learning in the lab loop
Choosing the next experiment to maximise information per assay.
Developability for biologics
Early flags for aggregation, immunogenicity and expression risk.
Single-cell disease atlases
Cell-type-resolved views of which cells drive disease.
Causal target inference
Separating disease drivers from downstream consequences.
Nucleic-acid design
Potency, stability and specificity for siRNA and mRNA.
Explore