Protocol standard
Each experiment is specified once, in layers, so that any qualified laboratory can run it the same way. Experimental variables remain open; procedural choices with no scientific rationale are standardized, because they create the batch effects that prevent data from being pooled.
Four layers
Aim
What to learn: the question, the variables, the outputs and the quality required. Never how.
Protocol
Standard operations (seed, treat, label, stain, measure) with built-in checks and what to do when a check fails: retry, flag or abort.
Calibration
The artifacts every lab must pass, and the context each measurement records: instrument, reagent lots, donor or passage, operator, uncertainty.
Instruments
The only layer specific to one laboratory's equipment. Changing laboratories changes this layer and nothing else.
Change proposals
A standardized setting changes only through a change proposal, with evidence that the change improves the science, reviewed by the laboratories that run it.
| Proposal | Title | Status | Summary |
|---|---|---|---|
CP-01 | Standardize neutral-lipid staining | accepted | BODIPY 493/503 at 1 µg/mL for 30 min after 4% PFA, 9 fields per well. Local variants caused lab-to-lab differences with no scientific reason. |
CP-02 | Standardize hepatocyte seeding and attachment | accepted | Collagen-I 96-well plates, 0.8 × 10^5 cells/cm², 24 h attachment. Density was the largest arbitrary source of variance. |
CP-03 | Should the lipogenic medium be standardized? | open | Proposal to standardize 25 mM glucose + 100 nM insulin. Objection: glucose and insulin are disease-state variables some questions must vary. For now an experimental variable, declared in the aim. |
CP-04 | Standardize adipocyte insulin-resistance induction | open | Proposal to standardize TNF-α (10 ng/mL, 24 h) for insulin-resistant fat cells so labs in different countries produce comparable lipolysis data. |