Genomarker Diligence
Independent review of computational claims.
For investors weighing a thesis, BD teams evaluating in-licensing data, and translational groups inheriting someone else's analysis. Genomarker Diligence answers one question — independently, and in writing: does this result survive scrutiny, and can it be re-run?
How an engagement runs
Claim inventory
Every computational claim in the target material is identified and mapped to the analyses, data, and assurance profiles that could support it.
Checks
Each claim's analysis is reconstructed as a canonical analysis request and run through the same checks as every Genomarker analysis — design, confounding, leakage, plan alignment.
Independent re-analysis
Every examined analysis is re-run on our system from the original data, and each run produces a signed record.
Verdict report
A claim-by-claim verdict backed by structured findings and the signed record of every analysis we ran.
Why independent
The producer of a result — human or AI platform — cannot credibly vouch for it. Diligence verdicts are issued under a published independence policy: the statement says exactly what was verified, and never more.
What you receive
- Verdict reportClaim by claim, signed
- Structured findingsMachine-readable, with the rule behind each
- Signed analysis recordsOne per analysis we run
- Evidence packagesPortable, offline-verifiable — in development, Q1 2027
- Replay attestationsIndependent re-execution, classified — in development, mid-2027
The evidence, not the reviewer, carries the claim.
Every analysis we run for a review is signed against published keys — you can check our work without trusting us.