How a Peptide Specification Limit Gets Set

A certificate reports a result and, next to it, a limit. The result is a measurement. The limit is a decision somebody made, and how that decision was made determines whether the pass means anything at all.

A specification is three things, not one

A complete specification has an attribute, a method and an acceptance criterion. Remove any of the three and it stops being a specification.

“Purity: not less than 98%” is not a specification. Purity by what — chromatographic area, mass balance, net peptide content? At what wavelength, on what column, over what gradient? A number without its method is an assertion, which is the same point made about grade names in crude, desalted and purified grades. The attribute and the method together define what is being limited; the criterion says where the line sits.

Where a limit actually comes from

Limits are set from one of four places, and they are not equally meaningful:

• A compendial monograph. For a substance with a pharmacopoeial entry, the limit and the method are both prescribed. Most research peptides have no monograph, so this route is unavailable for them.
• A safety or toxicological basis. Used for impurities where a quantity has an established significance — elemental impurities and residual solvents are the usual examples, discussed in elemental impurities and residual solvents.
• Process capability. The limit is set from what the manufacturing process reliably achieves — typically the observed lot history with a margin. This is the most common basis in practice, and it is a statement about the process rather than about fitness for anything.
• Commercial convention. Ninety-eight percent is a market expectation, not a derived number. A limit set this way is a promise, and it is honest as long as nobody claims it was derived.

The fourth is far more common in this market than the first three. That is not in itself a problem — a conventional limit consistently met is still useful information — but it should be read as what it is.

Why the method has to be capable of the limit

A limit tighter than the method can resolve is decorative. If a chromatographic method quantifies reliably down to 0.1%, a specification of “any single impurity not more than 0.05%” cannot be assessed by it, and reporting a pass against that limit is reporting a measurement that was not made. The thresholds involved are set out in limit of detection and limit of quantitation.

The same applies in the other direction. A specification of “not less than 98.0%” assessed by a method whose uncertainty is plus or minus one percent is a limit the method cannot decide, as described in measurement uncertainty. Setting the limit and validating the method are the same exercise.

What a specification is quietly silent about

A specification constrains only the attributes it names. Everything else is unconstrained, no matter how many decimal places the named attributes carry.

A peptide can meet a 98% chromatographic purity limit while containing 25% counter-ion by weight and 8% water, because neither is an attribute of the specification. That is the distinction drawn in net peptide content, and it is a gap in the specification rather than a failure of the material. Reading a certificate well means noticing the attributes that are absent, not just the values that are present.

Release limits and shelf-life limits

A well-constructed specification often carries two criteria for the same attribute: a tighter one applied at release and a wider one applied for the remainder of the assigned period.

The logic is that the material degrades predictably, so a lot released at the shelf-life limit would fail it a month later. The gap between the two is taken from the stability data described in what a stability protocol contains, and it is why the assigned date carries different meanings in different systems, as set out in retest dates and expiry dates.

A specification with one criterion for both moments is either conservative at release or optimistic at the end, and usually the latter.

Reported, not detected, and the space between

Three entries mean different things and are often written as though they were interchangeable:

• “Not detected” means below the detection threshold of the method used. It is a statement about the method as much as the sample.
• “Not more than X” means the measurement was made and fell below X.
• “Complies” means a criterion was assessed without the value being disclosed, which withholds the information the buyer would use to compare two lots.

Where the actual number exists, reporting it costs nothing and tells the reader more than the verdict does.

Reading a specification you did not write

The practical questions are short. Which attributes are constrained, and which obvious ones are not? Is each criterion paired with a method? Is the method plausibly capable of deciding that criterion? Is a value reported, or only a verdict? And is the limit stated as a derived figure or a conventional one?

None of these require a laboratory to answer, and they separate a report that carries information from one that carries reassurance — a distinction that also separates the document types compared in certificate of analysis, conformance and origin. The broader reading framework is in how to read a certificate of analysis, and the reason two honest certificates for the same material can still disagree is in why certificates disagree on purity.