Case Study 34.1 — Reading a Certificate of Analysis, Line by Line

What this case study is, and what it is not

Below is a [constructed teaching example] of a certificate of analysis. It is invented for this book. It does not reproduce any real document, does not describe any real batch, does not name any laboratory or supplier, and no result on it corresponds to any measurement anyone has performed. It was built to contain the specific instructive features this chapter needs — some entries done well, some done poorly, and some conspicuously absent.

And a limit on how to use it. Working through this document teaches you to read an analytical claim precisely. It does not teach you to evaluate whether a preparation is acceptable to obtain or use, because — per §34.9 — that question is not answerable from any document, however good. A reader who finishes this case study feeling equipped to assess a purchase has taken exactly the wrong lesson. The correct lesson is the opposite: watch how many of the questions that matter turn out to be unaddressable by the document, and notice that the number does not shrink as the document improves.

The document

════════════════════════════════════════════════════════════════════════════════
  CERTIFICATE OF ANALYSIS                          [constructed teaching example]
════════════════════════════════════════════════════════════════════════════════

  Product .................. Peptide X (synthetic)
  Lot ...................... 24-0817-B
  Quantity ................. 10 mg / vial
  Date of analysis ......... [date omitted]
  Storage .................. Store at -20 °C, protect from light

  ── TEST ─────────────── METHOD ──────────── RESULT ─────────── SPECIFICATION ──

  Appearance             Visual              White powder        White powder
  Identity               ESI-MS              1,419.5             Conforms
  Purity                 HPLC                99.1%               ≥ 98.0%
  Peptide content        —                   —                   —
  Water content          —                   —                   —
  Counterion             —                   —                   —
  Residual solvents      —                   —                   —
  Sterility              —                   —                   —
  Endotoxin              —                   —                   —

  ── ─────────────────────────────────────────────────────────────────────────
  Released by: Quality Department
  ════════════════════════════════════════════════════════════════════════════
  FOR LABORATORY RESEARCH USE ONLY. NOT FOR HUMAN OR VETERINARY USE.
  ════════════════════════════════════════════════════════════════════════════

Line by line

Product — "Peptide X (synthetic)"

What it establishes: the name the supplier applied to this material.

What it does not: anything analytical. A product name is a label, and §34.6's hierarchy of identity claims puts "it came in a container labeled X" at the bottom for a reason. Note also what is missing from the identification: no sequence is stated. A document identifying a peptide only by name has not said which molecule it means, and for compounds circulating under laboratory codes rather than assigned generic names (Chapter 1, §1.8), the name may not resolve to one sequence unambiguously.

Lot — "24-0817-B"

What it establishes: that the supplier operates a lot numbering scheme, and that every result below is a claim about this lot and no other.

What it does not: that your vial belongs to this lot. This is limitation 1 from §34.7 and it is the largest hole in the document. The lot number appears on the paper. Whether it also appears on your vial, whether the two match, whether the material in the vial actually came from the fill of that lot, and whether this document was generated for this lot or reused from another — none of these are established by anything the document contains.

Notice something structural here. The lot number is the hinge on which the whole document's relevance turns, and it is the one entry that carries no test result at all. Everything below is measured; this is asserted.

Quantity — "10 mg / vial"

What it establishes: what the supplier intends to be in each vial.

What it does not: 10 mg of what. This is §34.4's entire argument in one line. Gross powder weight and peptide content are different numbers, and this entry does not say which one it is. Note that the certificate contains no peptide content result to resolve the ambiguity, so the ambiguity stays open. It is not resolvable from the document.

Date of analysis — omitted

What it establishes: nothing, and the omission is itself informative.

Without a date there is no interval between testing and now, which means no way to think about degradation, and no way to check the document against a stability program — which, since none is mentioned, does not exist. Rating rule 5 in this book's system requires claims to be date-stamped and falsifiable. A document that resists both is not making a claim that can be wrong.

Storage — "-20 °C, protect from light"

What it establishes: the supplier's stated storage condition.

What it does not: that the material was kept that way. This is limitation 2 from §34.7 — the document is a photograph of a past state and is not updated by events. Storage instructions describe what should happen going forward. They are silent on what happened between the analysis and your possession: how long the material spent in transit, at what temperature, through how many freeze-thaw transitions.

There is also a quieter point. A storage condition stated without a stability study behind it is a guess dressed as a specification. A real expiry comes from a formal program in which material is held under defined conditions and tested at intervals to establish when it stops meeting its specifications. Nothing here indicates such a program.

Appearance — "White powder / White powder"

What it establishes: a genuine, if crude, test. Gross visual abnormalities — discoloration, particulates, a collapsed cake — would fail it. This is not nothing.

What it does not: anything chemical. Many wrong things are white powders.

Note the specification is well formed: a stated criterion, a stated result, and a pass. Structurally this is the most complete row on the document, which should be sobering, since it is also the least informative.

Identity — "ESI-MS / 1,419.5 / Conforms"

What it establishes: a real result. Something in the analyzed sample had a measured mass of 1,419.5, and the supplier states this conforms to expectation.

What it does not:

  • Which mass convention was used. Monoisotopic or average? For a peptide of this size the two differ by roughly a dalton, which is more than the difference between some substitutions (§34.2's 🧬 callout). The document does not say.
  • What the expected value was. "Conforms" is a conclusion. The calculated mass is not printed, so the arithmetic cannot be checked by the reader.
  • Sequence. ESI-MS alone gives mass, not order. Had MS/MS been performed, the document would say so, and the identity claim would move up two rungs of §34.6's hierarchy.
  • Stereochemistry. Structurally invisible to any mass measurement.
  • Whether this was the only species. A dominant peak means a dominant signal, not a dominant mass fraction.

So the identity row is a real level-2 claim presented in a format that reads like a level-4 one. That gap — between what was measured and what a reader infers — is where most of the persuasion in documents like this happens, and it usually requires no dishonesty from anyone.

Purity — "HPLC / 99.1% / ≥ 98.0%"

What it establishes: that under whatever conditions were used, 99.1% of the detected peak area fell in the target peak, against a stated acceptance criterion. This is a well-formed row: method, result, specification, pass.

What it does not:

  • The method. "HPLC" is a family, not a method. Reversed-phase, ion-exchange, and size-exclusion separate by different principles and see different impurities. No column, no gradient, no run time.
  • The detector and wavelength. This determines what was counted at all. At 280 nm, entire classes of species contribute nothing to either the numerator or the denominator.
  • Whether specificity was demonstrated. No indication the method was shown to separate the target from known related substances. Without that, co-elution is not excluded.
  • Deletion sequences. Per §34.3 and Chapter 32, a species differing from the target by one residue may not resolve at all — and if it does not, its area is counted inside the 99.1%.
  • Anything about mass. Area percent is not a mass fraction.

There is a subtle trap in the specification column worth naming. "≥ 98.0%" makes the row look rigorous, and in one sense it is: a criterion was set in advance and met. But a specification only constrains what it measures. A tight limit on a number that is blind to the most likely impurity is precision aimed slightly to the left of the target.

Peptide content, water, counterion, residual solvents — all absent

What their absence establishes: nothing was measured, so nothing is known.

This is limitation 3 from §34.7, and it is the part of the document that most rewards deliberate attention, because human attention naturally follows what is present. Four rows of dashes are easy to scroll past. They are, collectively, the answer to "how much peptide is in this vial," and the answer is unmeasured.

Combine this with the "10 mg / vial" ambiguity and the shape of the problem is clear. The document reports a very high purity figure and no content figure, so a reader carries the reassurance from the first into the space left by the second. Per §34.4, that inference does not hold, and the ❌ rating on it is one of this chapter's four.

Note that nothing here implies wrongdoing. A supplier of research-grade material has no particular reason to determine peptide content, because laboratory users working quantitatively determine it themselves. The absence is a consequence of what the product is for.

Sterility and endotoxin — absent

What their absence establishes: these hazards were not assessed.

Per §34.5 these are two separate tests for two separate hazards, and neither implies the other. Their absence here is entirely expected: endotoxin testing is routinely not performed outside regulated manufacture, because the material is not intended for injection into anything.

This is also where the "NOT FOR HUMAN USE" line at the bottom stops being a formality and becomes a description of the document's own scope. The certificate was not written to answer the question of whether the material is safe to inject. It was written to describe the chemical properties of a laboratory reagent, and it does that reasonably well.

Released by — "Quality Department"

What it establishes: that the document asserts an approval.

What it does not: who approved it, whether that person was qualified, whether any record exists behind the assertion, or whether the tests described were performed at all. This is limitation 4 from §34.7 — a document cannot establish its own authenticity. In regulated manufacture this is addressed not by making documents harder to forge but by embedding them in an auditable system, where an inspector can trace a result back to instrument data and to a named analyst, and where a legal entity bears responsibility.

The bottom line — "FOR LABORATORY RESEARCH USE ONLY"

What it establishes: the supplier's statement of intended use.

This line is frequently read as legal throat-clearing. It is better read as the document's own statement of its scope, and it is consistent with everything above: identity and purity assessed, content and safety-relevant testing absent. The document is internally coherent. It says what it tested and does not claim more. Every gap examined in this case study is a gap the document itself is honest about, by omission.

Which is the uncomfortable conclusion. This certificate is not a bad document. It is a reasonable one for a laboratory reagent. The failure mode is not that the document lies; it is that a reader asks it a question it was never written to answer, and the document's silence gets read as reassurance.

What a fuller document would add, and what it still would not fix

Suppose every dashed row were filled in — content assayed against a reference standard, water and counterion determined, residual solvents within limits, sterility and endotoxin tested, MS/MS sequence confirmation, purity by two orthogonal methods with specificity demonstrated, a date, a named analyst.

That would be a genuinely excellent characterization package, far beyond anything in the consumer peptide space, and every column would improve.

And the four limitations from §34.7 would be exactly where they were. The document would still describe a sample rather than your vial. It would still say nothing about handling after testing. It would still be silent on whatever it did not look for. And it would still be a file. Improving the first list does not touch the second, which is the structural finding of this case study and the reason §34.9 is the section this chapter exists for.

Discussion Questions

  1. The lot number is the entry on which the entire document's relevance depends, and it is the one entry carrying no test result. Explain why this asymmetry is unavoidable rather than a flaw in this particular document, and describe what would have to exist outside the document to close it.

  2. The "10 mg / vial" line and the absent peptide content row interact. Explain the interaction in your own words, and state what a reader is likely to conclude that the document does not support. Why does this outcome require no dishonesty from anyone?

  3. The identity row is described above as "a real level-2 claim presented in a format that reads like a level-4 one." Using the hierarchy in §34.6, explain what would have to be added to actually make it level 4, and name what would still be invisible even then.

  4. The purity specification "≥ 98.0%" is well formed — a criterion set in advance and met. Explain why a well-formed specification on a number that is blind to deletion sequences provides less reassurance than its form suggests. Connect this to Chapter 32's account of what goes wrong during solid-phase synthesis.

  5. This case study concludes that the certificate is "not a bad document." Defend that judgment, then argue against it. Where does the responsibility sit when a technically honest document is routinely read as answering a question it does not address — with the writer, the reader, or the context that brings them together?

  6. Take the final section seriously. If filling in every absent row leaves all four of §34.7's limitations exactly where they were, what follows about the strategy of seeking better documentation? Answer without proposing any alternative route to evaluating a preparation, because §34.9 holds that no such route exists — the question is what a person should conclude from that, not how to get around it.