95% of CNS drug candidates fail in clinical development.We de-risk your candidates in human neurons – not mouse – before the clinic.Across the genetic diversity you would otherwise only meet in clinical trials.

Talk to us How it works

We specialize in neurodevelopmental disordersEpilepsy and autism cohorts available now · schizophrenia and ADHD in development

>300consented patients to choose from

Human genetic diversity shows up as diverse neuronal activity, measured from patient-derived neurons in a dish. Coloured vertical lines in a row are the electrical activity of the subject’s neurons.

01 · The problem

Nineteen out of twenty CNS drug programmes never reach the market.

Preclinical evidence comes from animals and single cell lines. Patient diversity comes in clinical Phase 2 & 3. That gap is where CNS programmes fail.

Mouse ≠ human

Animal genetics does not have the human complexity. What works in a mouse rarely translates to humans.

One ≠ many

A single cell line reflects one subject. A clinical trial has dozens to hundreds, each with a distinct genetic profile.

Phase 2 is late

Today pharma learns in Phase 2 whether a candidate works. That answer costs $25–80M in Phase 2 alone, on top of the $20–30M already spent to get there, and the odds are still slim.

02 · Our approach

Human diversity, tested in a dish.

Clinical-Trial-in-a-Dish: the trial before the trial. Neurons from consented patients, each carrying that patient's full genome, tested as a cohort and read with functional assays.

Human

Neurons from patients, not mice. iPSC-derived, carrying each patient's full genome.

Diverse

Dozens of subjects per study, drawn from a biobank of hundreds. A population, not a single cell line.

Disease-relevant

Readouts that measure neuronal function and drug efficacy: electrophysiology, drug screens, transcriptional profiling.

03 · What we offer

Ship us your compounds. We build confidence into your go/no-go decisions.

Any stage. Same question: does it work in patients?

From target validation through hit-to-lead and lead-to-candidate to candidate selection, and for clinical programmes that need a second look.

What an answer could look like

One compound. 30 neuronal patient models.

Each bar is one patient model's response to the same compound. The spread is the information: how much of the effect is shared across patients and how much is patient-specific.

A single cell line or an animal model gives you one bar. Thirty bars tell you what one cannot: how large the effect is, how many patients share it, and how much of the spread is biology rather than noise.

Does it work in human neurons?
Your candidate tested on patient-derived neurons with disease-relevant functional assays and their primary readouts: neuronal activity, gene expression, drug response.
Is the effect consistent across genetic backgrounds?
One background tells you about one background. Thirty tell you whether the effect holds up across the genetic diversity a clinical trial will meet.
Can responders and non-responders be separated by genetics or clinical phenotype?
This is what turns compound validation into a clinical trial design: an enrichment hypothesis before the first patient is enrolled.

Available today for epilepsy and autism spectrum disorder. Schizophrenia and ADHD cohorts are in development.

We work with biotech and pharma at every stage from target validation to candidate selection, and with clinical-stage assets that need a second look.

For AI-driven drug discovery we provide the human ground truth: patient-derived functional data covering human genetic diversity to train, fine-tune and validate foundation models for biology and neuroscience via lab-in-the-loop.

Discuss your programme

04 · Who we are