How to Read and Verify a Peptide Certificate of Analysis

A certificate of analysis is the one document almost every peptide supplier will send you, and the one most buyers read least carefully. It arrives as a PDF or a photograph, it has a table with a purity figure in the high nineties, and it has a signature. Most of the time the conversation ends there. This guide is about what to do instead: how to read each line, what each line can and cannot prove, and how to confirm that the certificate describes the material you actually received rather than a lot that was made months earlier, or by someone else.

The approach is deliberately mechanical. You will not need a chemistry degree to apply it, but you will need to be willing to open a few public databases and compare numbers. Everything below is illustrated on a real, public lot record maintained by Pepti, the manufacturer that maintains this site, because it is a record anyone can open; the same method applies to any supplier's document.

What a certificate of analysis is, and what it is not

A certificate of analysis (COA) is a record of tests performed on a specific lot of material, listing for each test the acceptance criterion (the specification) and the measured outcome (the result). It is produced by whoever ran the tests: usually the manufacturer's own quality unit, sometimes a contract laboratory, occasionally both on separate documents. The regulatory origin of the format is worth knowing because it explains what a good one contains. In the United States, 21 CFR 211.194 requires laboratory records to include "complete data derived from all tests", the method used, the sample identification, the calculations and the identity of the people who performed and checked the work. 21 CFR 211.165 requires that each batch be tested against specifications before release. Those rules apply to regulated drug manufacturing, not to research-use materials, but they describe the information a certificate exists to carry, and a certificate that omits most of it is a weaker document whatever the market.

Three things a COA is not:

• It is not an identity document for the vial in your hand. A certificate describes a lot. The link between that lot and your vial is the lot number printed on the label, and nothing else. If the lot numbers do not match, or the label has no lot number, the certificate describes something, but not necessarily what you received.
• It is not independent evidence. A manufacturer's COA is the manufacturer's statement about its own material. It can be entirely accurate and still deserve to be checked, which is the subject of the guide on independent third-party testing.
• It is not a statement about things that were not tested. A COA that lists purity by HPLC and identity by MS says nothing about water content, counter-ion, residual solvents, net peptide content or microbial quality. The absence of a line is information too.

A Certificate of Conformance (CoC), which some suppliers send instead, is weaker still: it states that the lot meets specification without showing any results. There is nothing in a CoC to re-check. Read a CoC as an invitation to ask for the COA.

The header: the identifiers that must match

Before the test table, every certificate carries a block of identifiers. These are the lines that make the document checkable, and they are where most problems show up first.

Dates deserve a moment. A certificate whose test date is before its manufacture date is describing an impossibility, and one whose test date is the same day as a long chromatographic method was run, for a dozen lots, is at least worth a question. In Pepti's public library, the record for lot PEP260712-RT10 shows manufacture on 12 July 2026, testing on 21 July 2026 and a retest date of 11 July 2028: a plausible sequence that you can read straight off the record page.

The lot number format matters more than people expect. A manufacturer that encodes information in its lot numbers (a date, a product code) gives you a way to spot inconsistencies. Pepti's codes take the form PEP, a six-digit manufacture date, a hyphen, and a short product-and-strength suffix, so PEP260712-RT10 should correspond to a Retatrutide 10 mg lot made on 12 July 2026, and it does. A lot number that is just "001" or that changes format between documents tells you less, and a lot number that appears on another company's website is a warning covered in the guide on supplier legitimacy.

The test table: specification against result

The body of a certificate is a table with at least three columns: test, specification, result. Read it in that order, line by line, and ask of each line three questions. Is there a specification at all? Is the result a number or a word? Does the method appear?

A specification is the acceptance criterion: "≥ 98.0% by HPLC", "conforms to expected molecular weight", "white lyophilized powder". ICH Q6A, the international text on specifications, describes a specification as a list of tests, references to procedures and acceptance criteria, and that triad is what you are looking for. A result column that reads "pass" or "conforms" against a numeric specification is not wrong, but it hides the number, and the number is what lets you compare lots or compare suppliers. Prefer certificates that print the figure.

The method column is the one most often missing from research-peptide certificates. "Purity: 99.2%" without "by RP-HPLC, 214 nm, C18, gradient X" is a number without a context, and the context decides what the number means. The guide on HPLC purity and MS identity goes through why.

Typical lines on a peptide COA, in rough order of how often they appear:

1. Appearance (visual)
2. Purity by HPLC (area-percent, with or without the chromatogram attached)
3. Identity by mass spectrometry (expected mass against observed mass)
4. Peptide content or assay (less common; by amino acid analysis, nitrogen determination or UV against a standard)
5. Water content (Karl Fischer or loss on drying; uncommon on research-use certificates)
6. Counter-ion content (TFA or acetate; uncommon)
7. Residual solvents (uncommon)
8. Bacterial endotoxins or bioburden (appears on some certificates; relevant only to specific document sets)

Lines 1 to 3 are the minimum. A certificate with only lines 1 and 2 has not confirmed identity, and a certificate with only lines 1 and 3 has not measured purity. Lines 4 to 7 are the ones that distinguish a thorough document from a routine one; their absence is normal in the research-use market, and their presence is a point in a supplier's favor.

Reading the purity line

The purity figure on nearly every peptide certificate is an HPLC area-percent: the area of the main peak divided by the total area of all integrated peaks, at a stated detection wavelength. Three consequences follow, and all three matter when you read the number.

First, area-percent is relative. It tells you what fraction of the material the detector saw was the main component. It does not tell you how much peptide is in the vial, because water, counter-ions and salts do not absorb at 214 nm and are invisible to the method. A vial can be 99% pure by HPLC and contain 80% peptide by weight; both statements can be true at once. Preston and Phillips, writing in Analytical and Bioanalytical Chemistry, put it plainly: using UV peak areas to estimate percentage purity on a mass basis assumes absorbance is uniformly proportional to mass, so "the result may be only semi-quantitative".

Second, area-percent depends on the method. A short, steep gradient merges neighboring peaks into the main one and raises the figure; a longer, shallower gradient separates them and lowers it. The same vial can produce 99.5% on one method and 97.8% on another, and neither laboratory has lied. This is why the method line matters and why two suppliers' purity figures are not directly comparable unless the methods are.

Third, the figure is only as good as the integration. Without the chromatogram, you cannot see whether small peaks near the baseline were integrated, whether the main peak has a shoulder, or whether the run was long enough for late-eluting material to appear. A certificate that attaches the chromatogram, or links to it, is giving you the evidence; one that prints only the number is asking you to take it on trust.

Across the 79 lot records in Pepti's public library at the time of writing, the reported HPLC purity figures ranged from 98.4% to 99.8%, with one blend recorded as a textual "conforms". That is a normal range for well-made research peptides, and it is also exactly the range within which method differences, not material differences, can account for the variation between any two figures. Treat tenths of a percent as noise unless you know the methods were the same.

Reading the identity line

The identity line usually reads "Molecular weight (MS): expected 1419.5, found 1418.5" or just "conforms". Mass spectrometry confirms that a molecule of the expected mass is present in the sample. It does not confirm purity (a mass spectrum is not quantitative in this sense) and it cannot see impurities that have the same mass as the target, which includes diastereomers and some sequence isomers.

What you can check on the identity line without any instrument is the expected value. Look up the compound's CAS number on PubChem, read the computed molecular weight and the monoisotopic mass, and compare. The observed value should match one of them within roughly a dalton, and the certificate should say which convention it used. A mismatch of about one dalton per thousand usually means monoisotopic and average masses are being compared with each other; a mismatch of exactly one dalton often means a protonated ion mass has been reported as if it were the neutral molecule. Neither is a sign of fraud; both are signs the certificate could be clearer. A mismatch of many daltons, or an observed value for a different compound, is a different matter.

One pattern worth noticing when you read several certificates from the same supplier: if the observed mass is identical to the last decimal across every lot of a compound, ask whether the value is a measurement or the expected value copied into the result column. Pepti's own library shows 4732.16 Da for each of its retatrutide lots; the honest explanation is that the figure is the deconvoluted value reported to a fixed precision, and the way to confirm it is to open the PDF for the lot and look for the spectrum. The point of this guide is that you are entitled to ask, and a supplier who publishes records should expect the question.

Lines that are often missing

The research-peptide market has settled on a two-test certificate: purity by HPLC, identity by MS. Four further measurements change what the two-test certificate means, and it is reasonable to ask whether a supplier can provide any of them for a given lot.

Net peptide content. The proportion of the vial's weight that is peptide rather than water or counter-ion. Typical values for lyophilized synthetic peptides are well below 100%, and the figure is what you need if you intend to prepare solutions of a known concentration. It is measured by amino acid analysis, nitrogen determination or UV absorbance against a characterized standard.

Counter-ion. Peptides purified by reversed-phase HPLC with trifluoroacetic acid in the mobile phase are isolated as trifluoroacetate salts unless the counter-ion is exchanged. The certificate may state "acetate" or "TFA salt" or nothing at all. Counter-ion changes gross weight and can affect downstream use; the guide on purity and how it is tested covers it in more depth.

Water content. Lyophilized material retains residual water. Karl Fischer titration measures it. Its absence from a certificate is usual; its presence tells you the supplier weighs and tests more thoroughly than most.

Related substances. A single purity number collapses every impurity into "the rest". A related-substances line lists the individual impurity peaks by retention time and size, and sometimes by identity. For regulated drug substances, ICH Q3A(R2) sets thresholds at which impurities must be reported, identified and qualified, starting at 0.05% for reporting. Research-use certificates almost never go this far, but a supplier who can show the impurity profile on request is showing you the chromatogram behind the number.

Who tested it, and who signed it

Every certificate has an issuer. Read the issuer line as carefully as the results, because it decides how much independence the document carries.

A manufacturer's in-house certificate is produced by the party with the strongest interest in a favorable result. That does not make it wrong, and in regulated manufacturing it is exactly how release testing works, under controls designed to make falsification difficult. FDA's 2018 guidance on data integrity describes those controls, including the expectation that records are attributable, legible, contemporaneous, original and accurate, and that audit trails are reviewed. A research-use supplier is not held to that guidance, but you can borrow its questions: who generated the data, when, on which instrument, and could the record have been altered after the fact?

A contract laboratory's certificate carries more independence, provided three things are true: the laboratory is identifiable and reachable, the sample it received is the lot in question (which depends on who selected and shipped the sample), and the report can be confirmed on the laboratory's own website rather than only on the supplier's. The registry on this site lists only reports that meet the third condition, and the third-party testing guide explains how to meet the second yourself.

A signature on a PDF is a formality unless it is attached to a name and role. "QC Manager" with a scrawl is normal; a certificate with no issuer name, no company identifier and no signature at all is a document anyone could have produced.

Verifying the document in five checks

Reading the certificate tells you what it claims. Verifying it tells you whether to believe it. The following five checks can be done in a few minutes with a browser and the vial in front of you.

1. Match the lot number on the label to the certificate, character by character

Not the product name, not the strength: the lot number. If the label carries a QR code, scan it and note where it goes. A code that opens the exact lot record on the manufacturer's own domain is strong; a code that opens a generic page, an image, or a third-party host is weak; no code is neutral. Pepti's labels and record pages carry a QR code that resolves to the canonical record URL on its own domain (the record page says "scan to open this exact canonical COA record"), which is the pattern to look for.

2. Open the record where it lives, not where it was sent

If the supplier maintains a public library of lot records, find the lot there rather than relying on the PDF in your inbox. A document that exists in a public, dated, browsable place is harder to fabricate retrospectively than an attachment. If the supplier has no public library, ask whether the certificate can be confirmed by the issuing laboratory.

3. Cross-check the compound identifiers against an independent source

Take the CAS number and expected molecular weight from the certificate to PubChem and confirm that they describe the compound on the label. A certificate that cites the wrong CAS number for the compound is either careless or copied from another product's template; either way it is not evidence about this lot.

4. Check the dates for sequence and plausibility

Manufacture before test, test before shipment, retest date after both and consistent with the stated storage conditions. If you received the vial a year after the test date, the certificate still describes the lot as tested then; the retest date tells you how long the manufacturer stands behind the figures.

5. Look for the same record elsewhere, including places it should not be

Search the lot number in a search engine and in the Wayback Machine. A lot number that appears on the manufacturer's own site in an archived snapshot from before you ordered is consistent with a real lot. A lot number that appears on a second company's website, presented as that company's own lot, means one of the two companies is copying the other's documents; the guide on supplier legitimacy describes a real case of exactly this.

Worked example: one public lot record

Open the record for lot PEP260712-RT10 in Pepti's public library (the link is in the references). It is a short page, and that is the point: you can read all of it in a minute and compare every field with the method above.

The header gives product (Retatrutide), specification (10 mg), lot reference (PEP260712-RT10), CAS number (2381089-83-2), manufacture date (12 July 2026), test date (21 July 2026) and retest date (11 July 2028). The lot code decodes to the manufacture date, which matches. The CAS number, taken to PubChem, returns retatrutide. The dates are in sequence.

The results block gives an HPLC chromatographic purity of 99.25% and an MS molecular-weight result of 4732.16 Da, with a conclusion of "meets specification". The purity is within the range of the rest of the library. The molecular weight is close to the computed average mass for the compound, and the fact that the same value appears on the other retatrutide lots is the kind of thing this guide has told you to ask about: open the linked PDF and look at what was measured.

The record offers the full COA PDF, a QR code resolving to the record's own URL, and a note that independent local verification is welcome. What it does not offer, and what you would have to ask for, is water content, counter-ion, net peptide content and an impurity table. That is a fair summary of where a reasonably well-documented research-use lot sits in 2026: more than most, less than a regulated drug substance, and clear about which is which.

Red flags, and how much weight to give each

Not every oddity is fraud. The table below separates the signs that should stop a purchase from the ones that should prompt a question.

Maintained by Pepti, a Guangzhou peptide manufacturer; lot records referenced here are published at peptioem.com/coa. This guide is informational and does not sell or recommend specific vendors. Corrections: contact.