How to Read a Peptide COA (Certificate of Analysis): A Field-by-Field Guide

When a research compound is not sold through a pharmacy, the certificate of analysis is the closest thing you have to a label you can audit. It is the document that tells you whether the vial in your hand actually contains what the catalog page promised — and at what purity, with what impurities, and free of what contaminants. Yet most buyers never learn to read past the headline percentage, which is exactly the number a low-quality supplier can print without earning.

This is a field-by-field guide to reading a peptide COA the way an analytical chemist would: identity first, purity second, and physical and microbiological properties to close. It is written for research and educational purposes; the compounds discussed are supplied for academic and pre-clinical study and are not approved for human use.

What is a certificate of analysis (COA)?

A certificate of analysis is a single-lot test report. Its job is to state, with measurements rather than adjectives, what a specific production run of a compound actually is. The word that matters most is lot: a COA is only a COA if it is tied to a named batch. A generic spec sheet reused across every shipment describes an intention, not the vial you received.

Read that table as a checklist. Every line is a question the document should answer without you having to ask. A COA that answers only the purity line — and answers it with a number and nothing else — is the analytical equivalent of a receipt with the total scratched in by hand.

Who issues a peptide COA: in-house vs third-party

Every COA is produced by someone, and who that someone is changes how much the document is worth.

An in-house COA is generated by the manufacturer's own quality lab. It uses legitimate methods — the same HPLC and mass spectrometry an independent lab would use — but it carries an unavoidable conflict of interest: the party being evaluated is also the party reporting the grade. That does not make in-house numbers fraudulent. It makes them unaudited.

A third-party COA comes from an independent analytical lab that has no stake in the outcome. The methods are identical; the incentive is not. Independent testing is the only step that confirms a self-reported figure, which is why serious sourcing treats the manufacturer's COA as a starting point and an independent retest as the actual verdict. Our own standard requires re-running identity and purity at a lab the manufacturer does not use, across two separate lots — the full process is in how we vet a new manufacturer.

What sections are on a peptide COA?

A complete peptide COA is organized into a handful of standard sections. Not every vendor labels them identically, but the analytical content is consistent, and each section answers a different question.

The rest of this guide takes these one at a time, in the order a chemist reads them — because identity comes before purity for a reason: a purity percentage is meaningless if the molecule being purified is the wrong one.

How to read the identity section (mass spec confirmation)

Identity is confirmed by mass spectrometry, which measures the mass-to-charge ratio of the ionized molecule and, from it, the molecular weight ^[2]. The COA reports two numbers you compare directly: the theoretical (calculated) molecular weight of the peptide, and the observed mass from the instrument. For a correct molecule these agree to within a fraction of a percent.

Two techniques dominate. ESI-MS (electrospray ionization) tends to produce multiply-charged ions, so the raw trace shows a series of peaks that software deconvolutes into one molecular weight. MALDI-TOF typically shows the singly-charged molecular ion more directly. Either is valid; what you want on the COA is the observed value and, ideally, the trace behind it.

The identity check is a gate, not a footnote. If the observed mass does not match the peptide's known molecular weight, nothing downstream matters — you could have a 99%-pure sample of the wrong peptide. This is why a COA that reports purity but omits mass confirmation is structurally incomplete: it has skipped the one measurement that tells you what you are looking at.

How to read the purity section and HPLC chromatogram

Purity is measured by reversed-phase high-performance liquid chromatography (RP-HPLC) run against a validated method ^[1]. The sample is pushed through a column that separates molecules by how strongly they interact with the stationary phase; a detector records each component as it elutes, producing a chromatogram — a plot of detector signal against time, with a peak for every distinct species.

The number you have been trained to look for, the purity percentage, is almost always an area-percent: the area of the main peak divided by the total area of all peaks, times 100. That definition is the key to reading it critically. A 99% purity figure means the target peptide accounts for 99% of the total detected peak area — and the other 1% is one or more other peaks the same run detected but did not name.

So the chromatogram is more informative than the percentage it produces. On a good trace you see one dominant, sharp, symmetric main peak and a flat baseline. On a concerning trace you see the main peak flanked by several smaller peaks (related impurities — often truncated or deletion sequences from synthesis), a sloping or noisy baseline, or a main peak that is broad or shouldered, which can indicate co-eluting species hiding inside the "pure" number. A percentage with no chromatogram asks you to trust the arithmetic without seeing the data it was computed from.

What purity percentage matters for peptides?

There is a real benchmark, and there is a more important nuance.

The benchmark most quality suppliers hold for injectable-intent research peptides is ≥98%, and ≥99% is achievable for well-synthesized compounds. But the number in isolation is the least useful part of the COA. Under ICH Q3A, impurities in a drug substance are expected to be reported above a threshold near 0.05%, identified above roughly 0.1%, and qualified above roughly 0.15% ^[3]. Translate that to a COA: any impurity peak that represents a meaningful share of the total should be named, not left as an anonymous remainder.

This is the single most important reframing in reading a COA. Two lots can both read "99%" and be different products; the percentage tells you how much is not the target, and the impurity profile tells you what that fraction is. A characterized process-related impurity is a very different thing from an unidentified peak no one can name.

Water and acetate content: the salt and moisture lines

Lyophilized peptides are not delivered as pure dry peptide. They come as an acetate salt and they hold residual moisture, and both eat into the mass you think you bought.

Water content is measured by loss on drying (USP <731>) or Karl Fischer titration, and for a freeze-dried peptide it commonly runs a few percent by weight ^[6]. Acetate content is the counterion — during synthesis and purification the peptide is paired with acetate, and that salt is part of the net weight; it is quantified by ion chromatography or titration and can run into the low double digits by weight ^[7].

The practical consequence is arithmetic. If a vial's net content is a lyophilized cake, the actual peptide mass is net weight minus water minus counterion. Together those two fractions can account for 10–20% of a vial's weight. A COA that reports them lets you compute how much active compound you truly have; a COA that omits them quietly overstates the peptide mass — which downstream becomes an overstated concentration when the vial is reconstituted. This is why these unglamorous lines matter for anyone who cares that a stated dose reflects real peptide. (The reconstitution mechanics themselves are covered in how to reconstitute peptides with bacteriostatic water.)

Endotoxin and sterility: USP <85> and USP <71>

For material intended to be reconstituted and injected in a model system, two microbiological tests separate a characterized product from an unknown one.

Endotoxin testing detects bacterial pyrogens — fragments of Gram-negative bacterial cell walls that survive even when the bacteria themselves are dead. It is run as an LAL (Limulus amebocyte lysate) assay under USP <85>, and the COA reports a result against a stated limit, typically in endotoxin units per milligram or per vial ^[4]. Endotoxin is not something purity or sterility testing catches; it is its own measurement, and its absence from a COA for injectable-intent material is a real gap.

Sterility testing, USP <71>, checks for viable microbial contamination by membrane filtration or direct inoculation into growth media ^[5]. The COA reports "meets" or "conforms" against the named method. Sterility and endotoxin are complementary: sterility asks whether live organisms are present, endotoxin asks whether the residue of dead ones is — a sample can pass one and fail the other.

Not every research peptide COA carries both lines, and their absence is not automatically disqualifying for non-injectable applications. But for anything intended to go into a living system, they are the difference between a compound that has been checked for contaminants and one that simply has not.

Batch and lot numbers: tying the COA to your vial

Everything above is worthless if the document does not describe your vial. That link is the lot (batch) number, and it is the field most often faked by being left static.

A legitimate COA carries a lot number that is unique to one production run and printed identically on the vial label and the certificate. It should be accompanied by a manufacture date and often a retest or expiry date. When you receive a vial, the first cross-check is trivial and decisive: does the lot number on the glass match the lot number on the PDF? If a supplier sends the same COA — same lot number, same figures — for every order across months, they are handing you a template, not a test of your batch. The whole point of lot-specific documentation is that it fails this test honestly when the batch changes.

A worked example: reading a retatrutide COA row by row

Here is a realistic COA for a retatrutide lot, read the way you should read your own. The numbers below are illustrative teaching values, not a specific shipped lot.

Read it in order. First, identity: 4731.3 observed against 4731.5 theoretical is a match, so you know this is retatrutide and not a mislabeled lookalike. Only then does purity count for anything — and 99.1% is strong. Crucially, the impurity lines reconcile: the largest single impurity (0.4%) is identified, total impurities (0.9%) plus purity (99.1%) sum to 100%, and nothing is left as an anonymous remainder ^[3]. Then the physical lines — 4.2% water and 8.6% acetate — let you compute that roughly 87% of the net weight is actual peptide. Finally the microbiological lines confirm the material was checked for endotoxin and sterility. A COA that walks cleanly through that sequence is doing its whole job. The compound this example describes is the same one covered in the complete retatrutide guide.

How to spot a fake or edited COA

Once you know what a complete COA contains, the fakes become easy to spot — they are almost always defined by what they leave out.

• No lot number, or a frozen one. The most common tell. A certificate with no batch identifier, or the identical lot on every order, is a template.
• A number with no data behind it. A bold purity percentage and no HPLC chromatogram or mass-spec trace. The headline is the easiest thing to type and the hardest to fake with real data attached.
• A mass that does not match. Cross-check the reported molecular weight against the peptide's known value. A mismatch — or a missing identity section entirely — is disqualifying.
• Editing artifacts. Mismatched fonts, misaligned columns, a purity number in a slightly different typeface than the rest of the page, or a figure that looks pasted over another. These suggest a PDF was altered after issuance.
• Impurities that don't add up. If purity plus total impurities doesn't reconcile to 100%, or a "99%" figure sits above an unexplained remainder, the accounting is broken.
• No lab, method, date, or signature. A real COA names who tested it, against which method, on what date, with an analyst's authorization. Anonymous certificates are unaccountable by design.

None of these requires a lab of your own to catch. They require reading the document as a set of claims that must be internally consistent — which most edited COAs are not.

Why third-party testing matters more than the number

The recurring theme of every section above is the same: the value of a COA lives in the evidence, and the evidence is only as trustworthy as the party with nothing to gain from it. A manufacturer's own certificate can be entirely accurate and still cannot be audited on its own, because the grader and the graded are the same. Independent third-party testing removes that conflict — the same HPLC and mass-spec methods, run by a lab with no stake in the answer.

This is why the presence of a number is not the same as the verification of it, and why the cheapest source is so often the wrong one: independent testing across multiple lots costs real money, and suppliers who skip it can price lower while handing you a self-graded PDF. Our full standard — batch-matched COAs, independent retesting at a lab the manufacturer doesn't use, impurity-profile review, and two passing lots before a supplier is trusted — is documented in how we vet a new manufacturer, and it is the same lens to apply when deciding what to look for when buying retatrutide.

How to cross-verify a COA yourself

You do not need an analytical lab to do meaningful verification. A few checks, in order, catch most problems:

1. Match the lot number on the vial to the COA. If they differ or the lot never changes, stop here.
2. Check the mass. Compare the observed molecular weight on the COA to the peptide's known theoretical weight. They should agree to a fraction of a percent.
3. Reconcile the arithmetic. Purity plus total impurities should sum to 100%, with the largest impurities named.
4. Confirm the data exists. Look for an actual HPLC chromatogram and a mass-spec trace, not just typed numbers.
5. Read the physical lines. Water and acetate content should be present so you can compute real peptide mass.
6. Identify the lab. A named, dated, signed certificate — ideally from an independent lab — is auditable; an anonymous one is not.
7. Ask for an independent retest if the stakes justify it. A supplier confident in their material will not flinch at independent verification.

Run in that sequence, these steps take a few minutes and separate a document that earned its number from one that merely printed it.

For research and educational purposes only. Not medical advice. A certificate of analysis characterizes the contents of a specific lot; it does not make an investigational compound approved or safe for human use. The worked-example COA figures are illustrative teaching values. These compounds are supplied for academic and pre-clinical study and are not approved for human use.