Measuring the peptide inside the formulation

A collagen ingredient can pass a powder-based test without making the same test straightforward in a finished drink. Vitamins, plant compounds and other formulation ingredients may contribute signals that complicate measurement. A September study examines that problem for fish-derived collagen peptide beverages and proposes a two-step approach: identify interfering peaks, then use a better-matched calibration model.

The researchers compared three gel permeation chromatography methods associated with Chinese national or industry standards. Their beverage samples were two commercial products containing collagen and elastin peptides: one included vitamin C, while the other included epigallocatechin gallate, or EGCG, alongside additional ingredients. Peptide powders from the same production batches provided controls, helping separate formulation effects from differences in the starting material.

Interference comes before interpretation

The analysis used one HPLC system and a common column, with different mobile-phase compositions. Sample concentration was harmonised at 5 milligrams per millilitre, and measurements were performed in triplicate. This was a controlled comparison of analytical conditions, rather than a trial conducted across independent laboratories.

Under those conditions, vitamin C and EGCG eluted after 20.5 minutes. Excluding that late region helped align the beverage profiles with those of the matching peptide powders. The proposed cutoff is therefore tied to the tested setup and matrices. The authors recommend checking unfamiliar formulations with blank-matrix runs or spiking experiments before assuming that the same exclusion rule is suitable.

A curved calibration brings methods closer

The second issue was how retention time was converted into a molecular-weight estimate. A straight-line calibration left systematic differences between methods. Cubic fitting reduced those differences, with inter-method relative standard deviations below 5% for the combined fraction smaller than 1 kilodalton. That result matters when manufacturers compare specifications based on the proportion of small peptides.

Agreement was less complete when that fraction was split into narrower bands. Differences remained between one method and the other two in the 0.5–1 kDa and below-0.5 kDa ranges. The paper also used a 10% relative-deviation threshold, adapted from repeatability requirements, to classify differences between methods. Its use of the word significant should be read in that operational sense, rather than assumed to describe a conventional statistical hypothesis test.

A useful framework with a defined boundary

Electronic-nose and electronic-tongue measurements showed that the beverage formulations changed aroma and taste profiles relative to the powders. The corrected peptide analysis could still be performed in those matrices. This supports evaluating formulation and analytical performance together; it does not demonstrate consumer liking, improved absorption or clinical efficacy of the finished products.

The authors identify further validation needs: additional instrument platforms, other column batches and a wider range of beverage formulations. Resolution below 500 daltons is another limitation, for which more targeted methods such as LC–MS/MS could provide a cross-check. The immediate quality-control lesson is to record the matrix, separation conditions and calibration model together. A molecular-weight percentage without those details may conceal a methodological difference rather than an ingredient-quality difference.

SOURCES & TRANSPARENCYHow we work ↗
collagen peptidesbeveragesquality controlchromatography