A lab protocol is the exact, repeatable recipe for an experiment — every material, step, and parameter another scientist needs to get the same result. Here's what one contains, how it differs from an SOP, and why a protocol is only as good as its latest version.
A lab protocol is a precise, step-by-step description of how to carry out an experiment or procedure — listing the materials, reagents, equipment, exact parameters (time, temperature, concentration), and the ordered actions — so that another researcher can reproduce the method and obtain the same result.
A lab protocol is the executable form of a method. Where a paper's methods section summarises what was done, a protocol is detailed enough to actually do it: it names every reagent and its concentration, every instrument and its setting, and every step in the order you perform it. The test of a good protocol is simple — could a competent scientist who has never seen your experiment follow it and get your result?
That bar is higher than it sounds. The difference between "block the membrane" and "block in 5% BSA in TBST for 1 hour at room temperature" is the difference between a protocol that travels and one that only works in the hands that wrote it.
A complete protocol has a predictable anatomy. Most of these sections are obvious; the last one is the one almost everybody leaves out.
These three words get used interchangeably, but they solve different problems:
Most labs need both protocols and SOPs. The mistake is treating a protocol like a static SOP — freezing it in a document — when the whole value of a protocol is that it gets better as the science does.
A protocol is never finished. Someone swaps a buffer, corrects a blocking step, flags an antibody lot that ruined a week. The moment that happens, a static document becomes a liability: you now have two copies, and no record of which one is right or why the change was made.
This is exactly the problem version control for lab protocols solves. Every published change becomes a new version with a precise diff, tied to a reason and an author — so the method you cite next year is measurably more reproducible than the one you wrote today, instead of drifting into a folder of protocol_final_v3_FINAL.docx copies.
The reproducibility research is blunt about this: most published methods can't be reproduced by another lab on the first try, and most of what would help — the dead ends, the conditions that suppressed the signal, the steps that only work below a certain temperature — never gets written down. The path that worked is published; every wrong turn that mapped the way there is discarded.
That missing failure data is the highest-value part of a protocol, which is why jnar treats negative results as first-class data attached to the method. A protocol that records what doesn't work is a protocol the next person can actually trust.
It's the recipe for an experiment: the materials and equipment you need, the exact amounts and settings, and the ordered steps to follow. A good protocol is detailed enough that someone who has never run the experiment can reproduce it and get the same result.
A method is the general approach ("we quantified protein by Western blot"); a protocol is the specific, executable version of that method with every reagent, concentration, and step spelled out. The methods section of a paper summarises what was done; a protocol is detailed enough to actually do it.
An SOP (standard operating procedure) is an organisation's approved, often compliance-driven way of doing a routine task. A protocol is experiment-specific and evolves as the science does. SOPs optimise for consistency and audit; protocols optimise for reproducibility and improvement. Many labs run both.
A title and objective, scope and safety notes, a complete materials and reagents list with concentrations and catalog details, equipment, numbered single-action steps, the critical parameters (time, temperature, pH, speed), expected results, and — the part most protocols omit — what tends to go wrong and how it was fixed.
Because protocols change. A buffer gets swapped, a blocking step is corrected, an antibody lot is flagged. Without versions you can't tell which copy actually works or why a step is the way it is. Versioning ties every change to a reason and an author, so the method gets more reproducible over time instead of drifting.
