Peptifact

How to Read a Certificate of Analysis, and What One Cannot Tell You

A COA is usually the only document a research-peptide buyer ever sees. This page goes through it field by field, shows the arithmetic you can check yourself, and sets out what the paper does not certify.

Robert F · Published 2026-09-06

For most research-peptide purchases the certificate of analysis is the only document that ever changes hands. There is no label approved by a regulator, no package insert, no pharmacist. There is a PDF.

It is worth knowing what that PDF certifies, because it is both more specific and less reassuring than it is usually taken to be. This page goes through the fields, shows one check anyone can run with a calculator, and sets out the four things a certificate does not establish however good it looks.

This page describes documents and analytical methods. It does not tell anyone what to use or how much; our editorial standards explain where that line sits.

The fields, and what each one certifies

A research-peptide COA usually carries some subset of the following. What matters is which fields are present, because an absent field certifies nothing.

Field What it reports What it does not report
Product name and sequence The peptide the batch is claimed to be Whether the material in your vial is that batch
Lot or batch number Which production run was tested Anything about a run with a different number
Molecular weight The mass the analysis is checked against — this can be verified from the sequence, see below
Purity, RP-HPLC Share of the detected material that is the target peak Share of the vial's mass that is peptide
Peptide content Share of the vial's mass that is peptide Which impurities make up the remainder
Identity, mass spectrometry Whether the observed mass matches the expected mass Whether minor components are harmless
Water content Moisture, usually by Karl Fischer —
Counter-ion content Acetate or trifluoroacetate carried through purification —
Appearance, solubility Physical description Chemical identity
Endotoxin, sterility, bioburden Only if a test is listed Rarely present on research certificates at all

The distinction that does the most damage

Two numbers on a certificate look like they answer the same question and do not.

Purity by HPLC is an area percentage. The sample is separated by chromatography, the detector traces peaks, and purity is the target peak's area as a share of total peak area. It answers: of the material this method detected, how much was the peptide?

Peptide content is a mass fraction. It answers a different question: of the powder in this vial, how much is peptide? The remainder is water absorbed by a hygroscopic solid, the counter-ion the peptide was isolated with — acetate or trifluoroacetate — and residual salts. None of that is contamination in the sense of a failed synthesis; it is the normal state of a lyophilised peptide.

The consequence is direct. Purity and content can diverge, and a certificate that reports a purity figure and no content figure has not certified how much peptide is in the vial. That is the number every reconstitution calculation begins with. Our reconstitution page sets out the arithmetic — concentration equals peptide in the vial divided by diluent added — and if the first term is not what the label says, everything computed from it shifts by the same proportion.

A certificate reporting purity alone is not dishonest. It is answering a narrower question than most readers think they asked.

One check you can run yourself

A COA states a sequence and a molecular weight. Those two are not independent: the mass follows from the sequence. Anyone can verify that the document agrees with itself.

Add the standard average residue mass for each amino acid in the sequence, then add 18.02 Da for the water molecule released across the whole chain.

Worked, for BPC-157, sequence GEPPPGKPADDAGLV:

Residue Count Average residue mass (Da)
Gly (G) 3 57.05
Glu (E) 1 129.12
Pro (P) 4 97.12
Lys (K) 1 128.17
Ala (A) 2 71.08
Asp (D) 2 115.09
Leu (L) 1 113.16
Val (V) 1 99.13

Sum of residues 1,401.53, plus 18.02 for water, gives 1,419.55 Da. PubChem lists BPC-157 (CID 9941957, formula C62H98N16O22) at 1,419.5.

The same arithmetic on the three-residue peptide GHK gives 340.38 Da, and PubChem lists 340.38.

This does not tell you what is in the vial. It tells you whether the certificate is internally consistent — whether the sequence, the formula and the mass on the page describe the same molecule. A document that fails this check has a problem before any question of testing arises.

Four things a COA does not establish

That your vial is the batch tested. The certificate is bound to a lot number. If the vial carries no lot number, or a different one, the document describes something else.

That the laboratory is independent. A header is not provenance. The questions are whether the laboratory is named, whether it can be found to exist independently of the seller, and whether its certificates turn up across unrelated sellers in ways that do not make sense.

That the material is sterile or endotoxin-free. These are separate tests. If they are not listed, they were not reported, and a purity percentage says nothing about either.

That the impurities are harmless. A purity figure of any size leaves a remainder, and the certificate rarely says what the remainder is. This is the part that pharmaceutical regulation treats most seriously, and it is worth seeing why.

What the pharmaceutical standard looks like, for contrast

A synthetic peptide entering the US market as a generic medicine is held to something a research certificate does not attempt. FDA's framework for these products set limits on the relative abundance of each impurity and asked that any new impurity above a concentration threshold be assessed for its capacity to provoke a T-cell response, using methods that test both binding to human leukocyte antigens and the ability to drive T-cell proliferation.

That is not a theoretical precaution. Scientists at FDA's Center for Drug Evaluation and Research applied exactly those methods to teriparatide and its product-related impurities — in silico screening, HLA-DR binding assays, and a human peripheral blood mononuclear cell T-cell assay — and reported that the orthogonal approaches "identified multiple impurities as more immunogenic than TPT," the drug itself (Front Immunol 2025;16:1730346). The unnamed remainder on a certificate is the category that finding is about.

The framework is currently being rebuilt. On 2026-07-28 FDA published 17 revised draft product-specific guidances for peptide products — among them semaglutide, tirzepatide, liraglutide, teriparatide, glucagon, calcitonin salmon, dasiglucagon, pegcetacoplan and vosoritide — and withdrew the May 2021 guidance "ANDAs for Certain Highly Purified Synthetic Peptide Drug Products That Refer to Listed Drugs of rDNA Origin" on the stated ground that it "no longer reflects FDA's current scientific thinking." Anyone citing that 2021 document today is citing a withdrawn guidance.

None of this is a standard research-peptide sellers claim to meet. It is the measure of how much a single purity percentage leaves unanswered.

Where the certificate sits in the wider record

A COA is a chemistry document. It is silent on regulatory status, and the two are frequently confused. Products sold as research peptides are not approved medicines, and FDA has written to a series of sellers saying so — our warning-letter tracker records fifteen such letters between 2024-12-10 and 2026-08-24, in which the agency rejected research-use-only labelling where the website's own content showed the products were intended for people. A certificate of analysis does not change that analysis in either direction. It reports what a laboratory measured.

Sources and dates

Frequently asked questions

What does a certificate of analysis actually prove?

That a laboratory ran the listed tests on a sample of the stated batch and got the stated results. That is a narrower claim than it looks. It says nothing about a different batch, nothing about whether the vial received is from the batch named, nothing about handling or storage after testing, and nothing about any property that was not tested. A COA is a measurement record for one sample, and the value of the record depends entirely on who did the measuring.

What is the difference between purity and peptide content on a COA?

Purity by HPLC is the share of the detected material that is the target peptide — the main peak's area as a percentage of total peak area in the chromatogram. Peptide content is the share of the vial's mass that is peptide; the rest is water, the counter-ion the peptide was isolated with such as acetate or trifluoroacetate, and residual salts. They answer different questions. A vial can be highly pure by HPLC and still contain noticeably less peptide by mass than the label states, because the two numbers measure different things. If a COA reports only purity, the mass of peptide in the vial has not been certified.

Why does peptide content matter for reconstitution?

Because every reconstitution calculation starts from the milligrams of peptide in the vial, and that is the figure peptide content addresses. Our [reconstitution page](/how-to-reconstitute-peptides) sets out the arithmetic: concentration equals peptide in the vial divided by diluent added. If the vial's actual peptide mass is not the labelled mass, every number downstream of it moves by the same proportion. A COA reporting purity but not content leaves that starting figure unverified.

Can I check a COA myself?

Parts of it, yes, and one part is straightforward arithmetic. A COA states a sequence and a molecular weight, and the molecular weight follows from the sequence: add up the standard average residue masses for each amino acid and add 18.02 for one water molecule. BPC-157's sequence GEPPPGKPADDAGLV gives 1,419.55 Da, and PubChem lists BPC-157 at 1,419.5. If a certificate's stated mass does not match its stated sequence, something on the document is wrong. You can also check that the lot number on the certificate matches the lot on the vial, that the test date precedes the manufacture date sensibly, and that the laboratory named is an entity that exists and is not the seller.

Is a third-party COA better than the manufacturer's?

Only if the third party is genuinely independent and can be identified. A certificate naming a laboratory that has no traceable existence, or that is a related entity of the seller, is not third-party testing regardless of what the header says. The checkable questions are whether the laboratory is named at all, whether it can be found independently, whether the report carries a lot number that matches the product, and whether the same laboratory's certificates appear across unrelated sellers in ways that do not make sense.

How does a research-peptide COA compare with a pharmaceutical one?

They are not close. A synthetic peptide entering the US market as a generic medicine has had each impurity's relative abundance limited, and new impurities above a threshold assessed for the ability to provoke a T-cell response — an assessment FDA's own scientists have shown matters, having found multiple teriparatide impurities more immunogenic than teriparatide itself. FDA is currently rewriting that framework: it published 17 revised draft product-specific guidances for peptide products on 2026-07-28 and withdrew the 2021 guidance as no longer reflecting its current scientific thinking. A research COA reporting a single purity percentage is not attempting the same exercise.

Does a COA make a peptide safe to use in a person?

No, and no certificate claims to. A COA reports chemistry on a sample. Safety in a person depends on the identity and quantity of every impurity, on sterility and endotoxin, on the manufacturing conditions, and on evidence from clinical study — none of which a purity percentage addresses. Products sold as research peptides are not approved medicines, and FDA has written to sellers about exactly that; our [warning-letter tracker](/vendors/fda-warning-letters) records those letters.