Blog

Notes on reproducibility

Practical writing on protocol drift, versioning, and what it takes to make an experiment rerunnable.

ReproducibilityProtocol driftLab practice

The drift you never wrote down

Preclinical results fail to reproduce not because the science is wrong, but because the run that worked and the run that did not were never actually the same run.

Dr. Maya Ellison

July 5, 2026 · 5 min read

ReproducibilityProtocol versioningLab operations

Version your experiments like software

Treat every protocol as a versioned artifact. Capture it once, diff every change, and tie each result to the exact protocol version and reagent lots that produced it.

Daniel Okafor

July 16, 2026 · 6 min read