How to Read a Peptide Certificate of Analysis

The certificate of analysis is the only piece of paper that meaningfully separates a research-grade peptide from a labelled bottle. Marketing copy describes what a vendor wishes the lot were; the COA describes what the lot actually is, against documented analytical methods, signed by the laboratory that ran the tests. Reading one well takes five minutes, but only if you know which lines to read first and which absences to take seriously. This article is a field-by-field reading guide.

Two companion articles cover related ground. What ≥99% Purity Actually Means explains the math and method dependence behind the headline purity number; How We Verify Peptide Purity documents the QC pipeline a single vendor (PepMax) runs against. This piece is vendor- agnostic. The goal is to give a researcher evaluating anyCOA — ours, another vendor’s, or one inherited with a sample — the same field-by-field framework an accredited lab uses to read the same document.

Why a COA is the only document that matters

Anyone can write “≥99% pure” on a label. That phrase has no analytical content unless it is anchored to a method, a wavelength, an instrument, an analyst, and a date. The COA is the document that anchors it. Specifically, it does three things at once:

• It identifies the lot. A COA without a lot number that matches the vial in front of you is paperwork for a different lot. Lot numbers are the join key between the document and the physical material.
• It documents the method.“Purity 99.4%” is meaningless without “by RP-HPLC at 220 nm against a reference standard, system suitability checked.” The method line tells you what the number actually answers.
• It assigns responsibility. A signed COA is a statement by a named laboratory that the result is theirs. An unsigned PDF is anonymous, which is the same as anonymous in any other professional context: useful as a draft, not as a record.

Stores that ship a peptide with a glossy “quality summary” instead of the underlying COA are doing something specific: they are interposing themselves between you and the laboratory record. That is a choice. Treat it as a data point.

Anatomy of a complete COA

A complete certificate has roughly twelve fields, organized in four blocks: header, identity, purity and contaminant panel, and provenance. The exact layout varies by lab, but the blocks are universal. If a block is missing, the document is incomplete — it is not a stylistic choice.

Two things follow from this list. First, the COA is shorter than people expect — twelve fields fit on two pages. Second, every shrinkage from those twelve fields is information. A single-page summary with no chromatogram, no spectrum, no contaminant panel, and no signature is not a COA. It is a label.

Reading the HPLC chromatogram

The chromatogram is the most informative part of the document. The percentage typed underneath it is a summary; the trace itself is the underlying data. A researcher who learns to read the trace can tell in thirty seconds whether the number on top is plausible.

What every chromatogram should show

• A labelled y-axis from zero. Detector response in milli-Absorbance Units (mAU), starting at 0. A y-axis cropped to begin at, say, 50 mAU exaggerates the main peak and visually suppresses small impurities.
• A labelled time axis covering the full run.Typical reverse-phase peptide methods run 30–45 minutes including re-equilibration. A trace that ends at the main peak is showing you a window, not the run.
• A single sharp main peak.Tailing factor for the principal peak should be ≤2.0 under USP <621> criteria^[1]. A heavily tailed peak suggests overload or column degradation.
• Visible baseline and integration marks. The software-drawn integration start and end lines on each peak should be visible. Where they were drawn is part of the result.
• An integration table.Retention time, area, area %, and peak name for every peak above the integration threshold. The number reported as “purity” is the area % of the main peak in this table.
• Detection wavelength annotated. 220 nm is standard for peptide bonds; 210 nm is sometimes used as a more sensitive secondary trace. The number of percent purity is wavelength-dependent.
• A reference-standard or system-suitability annotation. Either a system-suitability run on the same analytical batch or a reference-standard injection tying the column performance on this run to a known control.

Red flags on the chromatogram

None of these are individually proof of fraud — a missing time axis can be a formatting accident — but their cumulative density on a document is informative. A well-run accredited lab’s output is consistent: every chromatogram in their library has the same axes, the same annotations, the same integration table format. Heterogeneity of formatting across lots from a single “lab” is itself a tell.

Reading the mass spectrum

Purity and identity are different questions. Purity asks how much of this lot is one molecule; identity asks which molecule it is. A high-purity lot of the wrong peptide is still a wrong-peptide problem. The mass-spectrometry field on the COA is the identity check.

For peptides below ~5 kDa, the standard method is electrospray ionization (ESI) on a quadrupole time-of-flight (Q-TOF) or Orbitrap instrument. ESI produces a charge-state envelope — multiple peaks at different charge states corresponding to the same molecule — that the software deconvolutes to a single intact mass. Three things should appear on the COA:

• The deconvoluted intact mass, observed.
• The theoretical monoisotopic mass calculated from the sequence.
• The mass error in parts per million (ppm) between observed and theoretical. Within ±5 ppm is the standard acceptance threshold; ±10 ppm is the loosest defensible tolerance for a Q-TOF result.

The spectrum itself should be attached: a charge-state envelope (multiple labelled peaks at consecutive charge states) is the underlying evidence. A single deconvoluted mass with no spectrum is the same kind of summary as a typed purity percentage with no chromatogram.

The five fields most stores hide

Beyond the headline purity number, five fields are most often quietly omitted from consumer-facing COAs. Each addresses something the HPLC chromatogram physically cannot see.

The pattern is consistent: each missing field corresponds to a contaminant or quality dimension that is not visible to the headline purity test. A vendor that wants to project rigor — without paying for it — will publish the HPLC purity number alone. That is a strictly weaker COA than one that publishes counter-ion, water, residual solvent, and endotoxin alongside the purity number, because it leaves the entire non-UV-absorbing contamination space unmeasured.

Numerical literacy: what the numbers actually mean

Several numbers on a COA look unambiguous and aren’t. Reading them well requires knowing which question each one actually answers.

• “≥99% pure” is a floor, not a value. A real lot might assay at 99.4%, 99.6%, 99.8%. The COA should print the actual measured value, not the marketing-floor specification. A document that gives only the floor is hiding the measurement.
• 97% vs. 99% is not a 2-percentage-point difference.The impurity budget triples: from 1% to 3%. For most research peptides that’s the difference between a single identifiable impurity and a heterogeneous bag of deletion sequences, oxidation products, and synthesis byproducts.
• Net peptide content ≠ HPLC purity.Net peptide content (sometimes “peptide content” or “peptide assay”) is the parent-peptide mass divided by total lyophilized mass, and it accounts for water and counter-ion. For a 99% HPLC-pure lot with 8% TFA and 5% water, net peptide content can be ~85%. Both numbers are correct; they answer different questions.
• Endotoxin in EU/mg, not just “passed.”“USP <85> compliant” with no number means nothing without knowing the limit applied. A research-tier limit (≤5 EU/mg) and a parenteral-tier limit (≤0.5 EU/kg/hr-of-use) are different by an order of magnitude.
• “Conforms to specification” is not a result.A line that says “Within Spec” with no measured value is a vendor-summary artifact, not a laboratory finding. The number was measured; printing it is the entire point.

Provenance, signature, and accreditation

The bottom block of the COA — lab name, accreditation, analyst signature, date — is the part that converts measurements into a record. Three checks live here.

• Lab letterhead, not vendor letterhead.The COA should be on the testing laboratory’s letterhead or template, not the vendor’s. A vendor-letterhead document with “tested by [Lab Name]” in small print is a vendor summary of a lab finding, not the lab’s own document. The lab’s own document is what was archived in their LIMS.
• Accreditation number, verifiable in five minutes. An ISO/IEC 17025- accredited lab carries an accreditation number from a national accreditation body. The three most common in peptide testing are ANAB (United States)^[7], A2LA (United States)^[8], and UKAS (United Kingdom)^[9]. Each maintains a public searchable directory. Type the number from the COA into the directory and verify (a) that the lab is currently accredited and (b) that the methods on the COA — chromatography, mass spectrometry, Karl Fischer, ion chromatography, GC-headspace, LAL — are on the lab’s accredited scope. Accreditation is method- specific, not company-wide.
• Analyst signature and date of issue. A named analyst signs; a date stamps the issue. An unsigned PDF, or a PDF with a stamped image of a signature and no name printed underneath, is not a signed document. The signature ties responsibility to a person with credentials, not to a vendor brand.

A red-flag checklist

The patterns below recur. None is individually conclusive; their cumulative density is.

A worked walkthrough

Suppose a vial of BPC-157 arrives with an attached PDF. Five minutes of reading is enough to grade the document. Run the checks in this order:

1. Lot match.The lot number on the vial matches the lot number on the COA, character for character. If not, stop — this is paperwork for a different batch.
2. Issuing laboratory.The COA is on the testing laboratory’s letterhead, or clearly identifies an independent lab as the issuer with the lab’s full address and accreditation block. The vendor’s name appears as “client,” not as issuer.
3. Accreditation verification.The COA shows an ISO/IEC 17025 accreditation number. Open the relevant accreditor’s public directory^[7][8][9], search the number, confirm the lab is currently accredited, and confirm that chromatography, mass spectrometry, and the contaminant-panel methods are on the lab’s accredited scope. This step takes five minutes and is the single highest- value check on the entire document.
4. Identity block. Sequence, theoretical mass, observed deconvoluted mass, ppm error. Mass spectrum attached as an image. ppm error within ±5.
5. Purity block. Chromatogram with full y-axis from zero, full time axis, integration table, detection wavelength stated, system-suitability or reference- standard annotation. Main-peak area % matches the headline purity number to one decimal.
6. Contaminant panel. Karl Fischer water (numeric), counter-ion content (numeric), residual solvents (Class 2/3 limits), endotoxin (numeric in EU/mg), bioburden where applicable.
7. Signature and date. Named analyst, signature, date of issue. Date of issue close in time to the manufacture date.

A document that survives all seven checks is a real COA. A document that fails one of them is informative. A document that fails three or more is a label.

For background on the analytical methods themselves — what HPLC at 220 nm physically measures, why ESI-Q-TOF is the identity standard for peptides under 5 kDa, what ISO 17025 accreditation does and does not guarantee — the two companion articles linked above cover the underlying chemistry. This piece’s contribution is the reading workflow: which fields to look at first, which absences to weight, and how to verify the accreditation claim against an external public source.