Every control answers one specific question, which means a control chosen without knowing its question establishes nothing. The commonest reason a peptide experiment cannot be interpreted is not that the measurement failed but that the comparison needed to make sense of it was never run. Controls are cheap at the design stage and impossible to add afterwards.
The four controls, and what each rules out
| Control | What it contains | What it rules out |
|---|---|---|
| Untreated | Nothing added | Baseline drift; tells you where the system sits undisturbed |
| Vehicle blank | Solvent at the same final concentration, no peptide | That the solvent produced the effect |
| Positive control | An agent known to produce the readout | That the assay was working at all |
| Negative or scrambled control | An inactive or sequence-shuffled peptide | That any peptide at that concentration would do it |
The untreated and vehicle controls are not interchangeable, and treating them as one is the most frequent shortcut. An untreated well tells you the baseline. A vehicle well tells you the baseline in the presence of the solvent, which for anything other than water is a different number. An organic co-solvent has effects of its own, and the only way to subtract them is to have measured them.
Why is a positive control worth a well?
Because without one, a negative result has two explanations and no way to choose between them: the peptide did nothing, or the assay did not work. Those are entirely different conclusions and they are indistinguishable from a flat plate. A positive control converts an uninformative null into a usable one — the system responded when it should have, and did not respond to the test compound.
This matters disproportionately in peptide work, where a peptide can vanish from the medium through proteolysis or adsorption before it ever reaches a receptor. A negative result from a degraded peptide looks exactly like a negative result from an inactive one. The positive control does not distinguish those two either, but it does eliminate the third possibility, and narrowing three explanations to two is real progress.
What does a scrambled peptide control establish?
That the effect depends on the sequence rather than on the presence of a peptide. A chain of the same residues in a shuffled order has the same mass, the same amino acid composition, the same charge and the same counter-ion load, and it should not engage the target. If it produces the same readout, the effect is a property of peptide material at that concentration rather than of the molecule — and that is a result worth knowing before building anything on it.
It is not always a practical control, since a scrambled version has to be synthesised to order. Where it is unavailable, an inactive analogue or a structurally related compound known not to engage the target serves a weaker version of the same purpose. The weaker control is substantially better than none.
Replicates: what kind, and why it decides what you can claim
Replicates come in two kinds that are routinely conflated, and the distinction determines what a statistic from them means.
Technical replicates are repeated measurements of the same sample — three wells from one preparation, three reads of one plate. They quantify the precision of your measurement. Averaging them gives a better estimate of that one sample's value.
Biological replicates are independent samples — separate cultures, separate animals, separate preparations from separate vials. They quantify the variability of the thing being studied.
Only the second supports a claim about the system. Three wells from one dilution of one stock describe one stock measured carefully; the error bar from them is a measurement error bar, and reporting it as though it described biological variability overstates the evidence considerably. The practical test is whether a replicate could have come out differently for a reason other than measurement noise — if not, it is technical.
Does a volumetric error count as biological variability?
No, and this is where the distinction becomes concrete for peptide work. Three wells pipetted from one tube share whatever error that tube carries — a stock prepared a few percent off its nominal concentration is a few percent off in all three. The replicates agree closely and all three are wrong by the same amount, which produces a tight error bar around a displaced mean. That is the worst available outcome, because tight agreement reads as reliability.
Independent preparations from separate vials are what break that correlation. The same applies to net peptide content, which varies lot to lot: an experiment repeated from a second lot is a genuinely independent test in a way that a second aliquot of the same stock is not.
How many replicates are enough?
That depends on the effect size and the variability, which is why the honest answer is that the number comes from the data rather than from convention. Three is a convention, not a calculation. What is worth saying plainly is that a concentration–response design extracts far more information from the same number of wells than a single-concentration comparison does, because a graded effect across a series is itself evidence that a difference between two points cannot provide. Spending a fixed number of wells on more concentrations rather than more repeats of one concentration is usually the better trade early on.
Do controls need to be run on the same plate?
Yes, wherever possible, because plate-to-plate variation is real and often larger than the effect being measured. Incubator position, edge effects, timing, and the age of the reagents all differ between plates run on different days. A control run last week cannot be subtracted from a treatment run today with any confidence. The experimental format sets limits on how much can be held constant, and within those limits the rule is to keep every comparison inside the smallest possible unit of variability — the same plate, the same session, the same operator.
The record that makes controls useful later
Controls only pay off if what was in them is recoverable. For each arm that means the peptide lot, the stock concentration and how it was derived, the vehicle and its final concentration, the passage number or animal cohort, and the date. This is a short list and it is routinely incomplete in exactly one place: the vehicle concentration, which is the field most often omitted and the one a reviewer is most likely to ask about.
Where material has been held for a while, storage history belongs in the record too, since a reconstituted stock has a horizon and an experiment run from a month-old aliquot is not the same experiment as one run from a fresh one. Keeping that history is ordinary good practice rather than a formality, and it is the difference between a discrepancy you can investigate and one you can only speculate about.
All products referenced here are supplied for laboratory and research use only. They are not drugs, foods, supplements or cosmetics, and are not for human or veterinary use.
