Why Two Laboratories Get Different Results From the Same Peptide

Two laboratories run what they believe is the same experiment on the same compound and get different answers. The usual first hypotheses are that one of them made a mistake or that one of the materials was substandard. Both are possible. Neither is the most likely explanation, because the number of things that had to match and were never specified is larger than most methods sections acknowledge.

Reproducibility is not one property

Three distinct questions get collapsed into one word:

• Repeatability — the same laboratory, same operator, same equipment, run again. Failure here is a problem with the method itself.
• Reproducibility — a different laboratory, different operator, different equipment, working from the published method. Failure here is usually a specification problem.
• Replication — a different laboratory asking the same biological question, not necessarily the same way. Failure here may be the most informative outcome of the three.

The distinction matters because the remedies differ. A repeatability failure is fixed by tightening the protocol; a reproducibility failure is usually fixed by writing down something that was never written down.

The material is the first place to look, and rarely the answer

Two lots of the same peptide are not identical, and several differences are invisible unless looked for. Net peptide content differs between lots, so equal weighed masses are not equal molar quantities, per net peptide content. The counter-ion may differ, which changes both the mass fraction and what else enters the well, per counter-ions and salt form.

Purity figures from two suppliers may not be comparable even when both are honest, for the reasons in why certificates disagree on purity. And where an impurity is biologically active, two lots of equal stated purity can behave differently because their impurity profiles differ.

These are genuine sources of divergence and they are the ones most often blamed, partly because the certificate is the only document in the chain and therefore the only thing available to blame. In most cases the larger differences lie downstream.

What happened between the vial and the well

By the time a compound reaches an assay it has been weighed, dissolved, aliquoted, frozen, thawed and diluted, and every one of those steps is a place where two laboratories diverge without either recording a deviation.

The weighing carries the hygroscopicity and static problems in weighing lyophilised peptides. The volumes carry the systematic error in pipetting accuracy — an uncalibrated instrument reading three percent low is invisible and consistent. Dilute working solutions lose material to surfaces, per adsorptive loss, and the loss depends on the labware each laboratory happens to use. Storage differs, including the unplanned cycling described in the auto-defrost problem.

None of these appear in a methods section. All of them change the quantity of intact compound actually present.

The biology that was never specified

The model contributes at least as much. Passage number, receptor expression level, serum lot, confluence and authentication status all vary between laboratories and are seldom fully reported — the territory covered in cell line choice.

Serum deserves separate mention because it is the least controlled reagent in most laboratories. It is a biological product that varies between lots in composition, in peptidase activity and in growth factor content, and a laboratory that changes serum lot mid-study has changed a variable nobody is tracking.

Why disagreement is not evidence of misconduct

Given that list, two competent laboratories differing is the expected outcome rather than a surprising one. The compound was nominally the same; the exposure, the model and the measurement were all slightly different, and effects in this literature are often modest enough that those differences are the same size as the effect.

There is also a publication asymmetry worth naming. Results that work get written up and results that do not usually do not, so the visible literature over-represents the conditions under which something was observed. A single positive report is therefore weaker evidence than it appears, which is part of the reading discipline in what the published literature actually shows.

What actually narrows the gap

Specify the material fully — supplier, lot, stated purity, net content, salt form — not just the compound name. Verify concentration rather than assuming it, by absorbance where the sequence allows, per concentration by A280. Keep one lot for a whole study where possible, so lot variation cannot masquerade as a finding. Record the preparation chain, not only the final concentration. And run a positive control with a known response, because a control that behaves as expected distinguishes a real negative result from a system that was not working.

The common thread is that most of what determines the answer is decided before the assay starts, and is not written down anywhere by default.

What a supplier can and cannot contribute

A supplier can make the material side traceable: consistent lot identification per batch and lot numbering, a certificate reporting values rather than verdicts, and clarity about what a stated mass refers to.

What no supplier can supply is the rest of the chain. The handling, the model and the measurement are assembled locally and account for the larger part of the variance — which is why the honest framing of a reproducibility problem starts with the vial and rarely ends there.