06 Identity & adulteration

How we prove a marine collagen lot is collagen, and from fish

Technical document 06 · Reviewed 2026-11-20 · Marine Collagen

On this page
  1. The four identity layers
  2. Hydroxyproline against protein, and the peptide balance
  3. What diffraction adds, and what it does not
  4. What adulteration looks like
  5. Writing the identity into the specification
  6. Five questions for the supplier

Three questions decide a marine collagen lot: is it collagen, has it been cut into peptides, and is it fish. The first is answered by hydroxyproline and the amino acid profile, the second by SEC-HPLC, and the third by a species test or a documented chain. The grade — ≥80% peptide, at least 80% below 3,000 Da, 80 mesh — is a fourth check that the certificate carries.

In one line

Collagen identity needs hydroxyproline, peptide identity needs the SEC curve, and source identity needs a species record; the protein figure proves none of the three on its own.

The four identity layers

Each layer answers a different question, and none of them substitutes for another. A certificate can be right on one and silent on the next, which is how a lot with the correct protein figure reaches a drum without anyone asking what the protein is or where it came from.

Identity checks · question, evidence and what each misses
LayerQuestionEvidenceBlind spot
CompositionIs it collagen?Hydroxyproline; amino acid profile with glycine, proline and hydroxyprolineSpecies and tissue
SizeIs it hydrolysed?SEC peptide content and the fraction below 3,000 DaSpecies; raw-material history
SourceIs it fish?Species PCR, or a documented species chainWhich tissue; whether the test sample matches the drum
GradeIs it the specified powder?Peptide content, ash, moisture, 80 meshSpecies; the raw material behind the lot

Hydroxyproline against protein, and the peptide balance

Collagen is recognised by hydroxyproline, an amino acid made by hydroxylating proline once the chain is built; other proteins carry little of it. A collagen peptide usually carries hydroxyproline at roughly 8–13% of protein mass. This grade does not fix a hydroxyproline limit, so the figure is an extra check rather than a release line, but it is the simplest identity screen available.

Two failures show up in the hydroxyproline-to-protein ratio. A non-collagen protein — soy, whey, pea or a hydrolysed cereal — raises the nitrogen without raising the hydroxyproline. Nitrogen spiking with urea or an ammonium salt raises Kjeldahl while the hydroxyproline stays where the collagen put it. The protein line alone shows neither.

The second ratio is the peptide balance. The catalogue asks for ≥90.0% protein and ≥80.0% peptide by SEC, with at least 80% of the peptide below 3,000 Da. Protein minus peptide is not one substance; it covers free amino acids, non-peptide nitrogen and method bias, so the SEC fractions are more informative than the difference. A certificate that reports only the protein figure has not shown how much of the powder behaves as peptide.

Keep the method beside the number. HPLC and colorimetric hydroxyproline do not return identical results, and a ratio built from two methods on one sample is stronger than one built across two certificates.

What diffraction adds, and what it does not

X-ray diffraction is a tool for crystals. It separates polymorphs and hydrates and produces a pattern that can be compared with a reference. A collagen peptide powder is amorphous, so diffraction returns a broad halo: it cannot say fish, and it cannot measure peptide size.

It still answers one useful question. A powder that should be amorphous but shows sharp crystalline peaks is carrying a crystalline carrier or a mineral residue — lactose, dextrose, sodium chloride or a phosphate. That is a mass-balance check, read beside ash and the SEC curve, not a species check. XRD is not a routine release test here, and a “conforms” line without the pattern is thin evidence. The XRD method note sets out what the technique can show.

What adulteration looks like

Substitution routes · the change and the test
RouteEffectDetection
Bovine or porcine collagenSpecies, while the collagen chemistry stays intactSpecies PCR; documented species chain
GelatinA high-molecular-weight tail; gelling on coolingSEC above 3,000 Da; gel test
Non-collagen proteinProtein up, hydroxyproline downHydroxyproline-to-protein ratio; amino acid profile
Nitrogen spikingKjeldahl up, peptide mass flatRatio; total nitrogen against amino-acid nitrogen
Crystalline carrierAsh or moisture; sharp XRD peaksXRD; ash; mass balance
Cheaper fish speciesThe species claim, not the collagen identitySpecies PCR; traceability record

No one test covers the list. The defensible file combines hydroxyproline, the amino acid profile, SEC and a species method, and that panel is modest for the risk.

Writing the identity into the specification

  • “Hydrolysed marine collagen, type I, fish skin; species and country of origin named.”
  • “Protein ≥90.0% by Kjeldahl with N × 5.55; factor and basis printed.”
  • “Peptide content ≥80.0% by SEC-HPLC; ≥80% below 3,000 Da; distribution bins reported.”
  • “Loss on drying ≤7.0%; ash ≤5.0%; ≥95% through 80 mesh.”
  • “Hydroxyproline result reported with its method, as an identity check.”
  • “Fish allergen statement; no other protein above the agreed limit.”
  • “CAS 9007-34-5 names collagen as a class; it does not name the species.”

That CAS line carries no source information: 9007-34-5 covers fish skin, fish scale and bovine hide alike, so a specification built on it alone says nothing about where the collagen came from. The COA page maps each figure to its place on the certificate, and the grades page explains the peptide and mesh gates.

Five questions for the supplier

  1. Which species and tissue is this lot made from, and does the COA name them?
  2. Which nitrogen factor and basis were used, and what is the hydroxyproline result and method?
  3. Can you provide the SEC fractions as well as the 3,000 Da cut-off?
  4. What supports the fish claim: a species PCR result, a documented species chain, or both?
  5. What is the fish allergen statement, and does the drum lot number match the COA and the retained sample?

Those five answers make a file; one protein percentage does not. Catalog ranges are listed on the product hub; the neighbouring fish grade is on the fish collagen peptides page; the methods library covers SEC and the other techniques; the quality page describes the retained sample.

Send the specification you intend to use — grade, quantity and destination market — and we will reply with the batch COA, the technical data sheet and a quotation.

Request a quotation