How CohortLayer validates a drug target
CohortLayer validates a drug target in three steps, without the underlying genetic data ever leaving the cohort. You define a target or genetic hypothesis; we test it inside the secure environments of two large human cohorts, separately; and you receive aggregate evidence — effect size, replication status, and phenotype and safety signals — that you can take straight into a decision. Because the two cohorts are tested independently and only the aggregate results are combined, a result that comes back has already survived a second, independent test. And because the evidence is bound to the cohort continuously rather than once, your target's evidence stays current as variants are reclassified and new findings are published. At no point is individual-level data downloaded, moved, or re-identified — only aggregate statistics are returned.
Three steps from question to evidence.
- 01
You define the question
A target, a variant, or a gene–disease hypothesis. We help shape it into something testable against population-scale data — the right phenotype definition, the right comparison, the right question to ask the cohort. - 02
We test it where the data lives
The analysis runs inside each cohort's secure environment, across two large human populations, separately. - 03
You receive aggregate evidence
Effect size, replication status, carrier counts, and phenotype and safety signals — returned as a clear, aggregate result. Keep the target monitored, and we tell you when the evidence shifts.
What are the steps to validate a target with CohortLayer?
There are three. Each is high-level by design — the method is precise, but the value is in the result, not the mechanics.
Step 1 — You define the question. A target, a variant, or a gene–disease hypothesis. We help shape it into something testable against population-scale data — the right phenotype definition, the right comparison, the right question to ask the cohort.
Step 2 — We test it where the data lives. The analysis runs inside each cohort's secure environment, across two large human populations, separately.
Step 3 — You receive aggregate evidence. Effect size, replication status, carrier counts, and phenotype and safety signals — returned as a clear, aggregate result. Keep the target monitored, and we tell you when the evidence shifts.
Why does CohortLayer test two cohorts separately instead of pooling them?
Because separate replication is a stronger test than a single larger dataset. A genetic signal that appears in one population does not always appear in another; pooling everything into one analysis can hide that. By testing each cohort independently and only then comparing the aggregates, a result that holds in both has already passed a replication check — the strongest filter against a false signal before you commit capital. Only the aggregate outputs are ever combined; the individual data stays separate and in place.
What does "the data never leaves the cohort" actually mean?
It means the analysis is brought to the data, not the data to the analysis. Managed-access human cohorts are governed environments — individual genetic records are not exported. CohortLayer runs the validation inside those environments and returns only aggregate statistics: counts, effect sizes, and signals that describe the population, never a person. Nothing is re-identified, downloaded, or moved. This is both a legal requirement for working with these cohorts and the core of our EU-native, aggregate-only design.
What do you receive at the end?
A clear, aggregate result you can act on: whether the association holds, how strong it is, whether it replicated across both cohorts, and what the phenotype and safety picture looks like — plus ongoing alerts when the evidence changes. You do not receive individual-level data, and you do not need to build or run any infrastructure yourself.
How long does the evidence stay valid?
As long as you keep the target monitored, it stays current. Genetic evidence is not static — variants are reclassified and new findings are published continuously, and the predictive value of human genetics is growing rather than saturating. CohortLayer re-runs the validation as the underlying evidence changes and flags what moved, so a target you cleared months ago is re-checked against today's evidence rather than frozen at the moment of the first report.
Frequently asked
Do you download or store our genetic data?
No. The analysis runs inside each cohort's secure environment, and only aggregate results leave.Is my hypothesis kept confidential?
Yes. Your target is your IP. We test it privately and the results come back only to you.How is this different from querying a public database?
Public databases report what has been published in general. CohortLayer binds that evidence to real cohorts, replicates it across two populations separately, keeps it continuously updated, and tests your specific private hypothesis — then reports what it means for your target.Is this a medical or diagnostic service?
No. CohortLayer returns aggregate, population-level evidence to organizations. It does not provide individual medical or diagnostic results.