Single Vial vs Batch Representative Testing

Almost everything written about peptide testing concerns instruments. Which chromatographic method, which detection wavelength, which mass spectrometer, how many decimal places.

The instrument is rarely the weak link. The weak link is upstream of it, in a step that produces no data and appears nowhere on the certificate: how the tested material was selected. Sampling design determines what a result can be generalized to, and no amount of analytical rigor downstream can widen a conclusion that the sampling made narrow.

What a result actually licenses you to say

An analytical result is a statement about the material that went into the instrument. That is the literal scope, and every honest certificate says so in its limiting language: results pertain to the sample submitted.

Everything beyond that is inference. When a supplier presents a certificate as evidence about the product on the shelf, an argument is being made that runs roughly like this:

1. The laboratory measured the sample accurately.
2. The sample is representative of the batch.
3. Therefore the batch is as described.

The laboratory is responsible for step one. Step two is where the whole chain either holds or does not, and it happens before the material ever leaves the client's facility. It is invisible on the certificate. It is also, in practice, where almost all the uncertainty lives.

The single-vial model

The dominant model in consumer peptide testing is straightforward: one vial goes to a laboratory, the laboratory tests it, a result comes back.

This model does something valuable. It answers the question a person actually has when they are holding a vial, which is what is in this. That is a real question, it has an urgent practical character, and single-vial testing is the correct tool for it.

The limits are just as concrete. A single result describes a single unit. Whether it describes the units next to it on the shelf depends on facts nobody in the transaction can observe: whether the run was homogeneous, whether the vial was drawn at random or selected, and whether the material tested came from the same run as the material sold.

That last point deserves emphasis, because it is the failure mode that all the analytical rigor in the world cannot touch. If the party submitting the sample also chose which sample to submit, and had any interest in the outcome, the sampling step is unblinded and self-selected. A laboratory receiving that vial will test it perfectly and report exactly what it found, and the result will be an accurate measurement of a sample whose relationship to the batch is unknown.

None of this is an argument against single-vial testing. It is an argument against reading a single-vial result as a batch statement.

What representative sampling means

Representative sampling is the practice of drawing material so that its composition reflects the population it came from. It is old, unglamorous and well developed, because entire industries depend on being able to say something about a lot without testing every unit in it.

The core requirements are consistent across those industries.

The population is defined. There is a batch, with boundaries, an identifier and a known size. Without that there is nothing to be representative of.

Selection is systematic rather than convenient. Units are drawn according to a rule set before anyone looked at them, not chosen by whoever packed the box.

Sampling spans the run. Production is not always homogeneous. Fill weight can drift, a lyophilization cycle can vary across a shelf, and a blend can separate. Material drawn from across a run detects variation that material drawn from one point cannot.

The sample size bears a defensible relationship to the batch. One unit out of ten thousand supports a weak inference regardless of how well it was analyzed.

Note what none of these are about. None of them is a laboratory function. They are all decided before the shipment is packed, which is why sampling design is a property of the testing engagement rather than of the analysis.

Why our minimum submission is 1,000 units

Novagen works with manufacturers and suppliers, not individuals, and testing engagements begin at a minimum of 1,000 units per submission. We do not offer single-vial consumer testing.

That threshold is not a commercial filter dressed up as a scientific one. It follows directly from what we are willing to have a certificate mean.

A submission at that scale is drawn from an actual production run, which means the batch exists as a defined population with an identifier. It allows material to be taken across the run rather than from one point in it. It gives a defensible relationship between what was tested and what the certificate is being used to describe. And it means the client submitting is the party who made or holds the batch, so the certificate attaches to a production lot rather than floating free.

The consequence is a narrower business and a certificate that means what it says. A result on a batch-representative submission can be described as a statement about that batch. A result on one vial cannot, and issuing a certificate that implies otherwise would be a documentation problem rather than an analytical one.

Two models, two questions

It is worth laying these out side by side, because they are frequently discussed as if one were a better version of the other. They are not. They answer different questions.

Single-vial testing answers: what is in this specific unit. Fast, accessible to an individual, directly relevant to the item in hand. Cannot support a claim about a batch.

Batch-representative testing answers: is this production run what it claims to be, consistently. Requires access to the batch, so in practice it requires the manufacturer or supplier to commission it. Supports a claim about the lot. Says nothing about a specific vial that has since travelled through a supply chain and been stored under unknown conditions.

A buyer holding a certificate benefits from knowing which of these produced it, because that determines what the document can be used to conclude. A supplier deciding what to commission should be clear about which claim they intend to make.

Questions worth asking about any certificate

• How many units did this result come from? The certificate rarely says, and the answer changes what the result supports.
• Who selected the sample? If it was the party with an interest in the outcome, the selection step was unblinded.
• Is there a batch identifier, and does it match the vial? Without that link the certificate cannot be tied to physical material at all.
• Was material drawn from across the run or from one point? This is what separates a batch statement from a spot check.
• Is the client on the certificate the party who made the batch? A certificate issued to a different company describes a different transaction. Our note on verifying a certificate covers why the client name is the check most people skip.

The honest position

No sampling scheme makes a certificate a guarantee about a specific vial that a customer eventually opens. Between the batch being tested and that vial being used sit packing, freight, customs, warehousing, a retail chain and a domestic refrigerator, none of which the laboratory saw.

What batch-representative testing does is remove one specific and avoidable unknown: whether the tested material had any defensible relationship to the material sold. That is a narrower claim than the market usually makes, and it is a claim that survives being examined.

Read next: how a submission works in practice, or how to choose a peptide testing laboratory.