Our Platform

Confidence in what biology to prosecute.

What is braid

Braid is Relation's platform. It takes our core goals—resolving disease biology, finding compelling assets and defining the clinical path—and breaks them into five targeted questions. Each question is a product, and each product is built from the models and data below.

Goal 1 Resolving disease biology

Braid genetics

What genes are causal?

Translates human genetic association into disease-relevant biology. A DNA foundation model and specialised insight models identify the genes behind a signal, predict regulatory effects and show the cellular contexts where they matter.

Runs on: DNA foundation models
Reads: GWAS, proprietary omics
Analyses: Statistical genetics

Why it matters

Higher-confidence targets

Targets backed by causal human genetics are far more likely to succeed in the clinic. Knowing which genes are causal raises confidence in every target we choose.

Braid mechanism

How does the target act in patients?

Turns candidate targets into testable hypotheses. Reasoning models organise proposed mechanisms into claims linked to evidence, grounded in literature, patient atlases and experimental results.

Runs on: Reasoning models
Reads: Literature, proprietary omics, wetlab
Analyses: Network and pathway analysis

Why it matters

De-risked hypotheses

Knowing how a target acts in patients tells us which biomarkers, modalities and patients to pursue, and exposes weak mechanisms before they cost a programme.

Braid cell

How does a cell respond to an intervention?

Predicts how human cells respond to perturbation. Models trained on single-cell and perturbation screens let us test interventions in silico before committing to the bench.

Runs on: Perturbation models, imaging models
Reads: Wetlab, proprietary omics

Why it matters

Fewer, better experiments

Predicting the effect of a perturbation in a patient is the grand challenge of drug discovery.

Goal 2 Find compelling assets

Braid assets

What targets, and in what indications, should we pursue?

Reads the clinical, commercial and scientific record continuously and returns a ranked, evidence-traced shortlist of external assets, then runs biology-first diligence on each one.

Runs on: DNA foundation models, reasoning models
Reads: Clinical data, literature, GWAS, proprietary omics
Analyses: Statistical genetics, patient stratification

Why it matters

Stronger portfolio bets

A ranked, evidence-traced shortlist means we invest in assets with real biological support, not just commercial momentum.

Goal 3 Define the clinical path

Braid trials

How could we improve our clinical trial outcomes?

Stratifies patients by mechanistic subtype and uses imaging-based models to enrich trials for the patients most likely to progress, so trials are smaller and more likely to read out.

Runs on: DNA foundation models, reasoning models, imaging models
Reads: Clinical data, proprietary omics, GWAS
Analyses: Patient stratification

Why it matters

Smaller trials, clearer readouts

Enrolling the patients most likely to progress raises the chance a trial reads out, with fewer patients and less time.

Our pipeline

Medicines that matter, developed to clinical value inflection points.

View our pipeline