Peptide Batch to Batch Comparability: Why Two Lots at 99 Percent Are Not the Same Material

Peptide batch to batch comparability is the question of whether two lots carrying the same name and the same purity figure are actually the same material. They frequently are not. A certificate of analysis describes one lot at one moment, and nothing on that document tells a researcher whether the synthesis route, the resin, the coupling reagents, or the manufacturing site changed between the lot tested last quarter and the vial on the bench today.

This is the quietest failure mode in research peptide supply. Contamination gets discussed. Purity gets argued about. Comparability almost never does, because the arithmetic looks reassuring: last lot reported 99.1 percent, this lot reports 99.2 percent, so the material must be equivalent. The number is a summary statistic over a chromatogram, and two very different chromatograms can produce the same summary. What changes between lots is the identity and proportion of the minor peaks, and those are precisely the features a headline purity percentage discards.

Why a purity number cannot detect a process change

Reverse phase HPLC area percent reports the fraction of total detected peak area attributable to the main peak. It is silent about what constitutes the remaining fraction. A lot at 99 percent may carry that residual one percent as a single deletion sequence eluting close to the main peak, or as fifteen unrelated species spread across the run, or as one racemised diastereomer that co-elutes and is not resolved at all. All three scenarios print the same figure.

Now change something upstream. Swap the coupling reagent, change the resin loading, move from a batch process to a continuous flow synthesiser, or transfer production from one contract facility to another. The main peak still integrates to roughly the same area because the chemistry still makes the target sequence efficiently. The impurity fingerprint underneath it shifts. Different truncations dominate. New oxidation products appear. Counterion content moves. None of that is visible in a percentage, and it is not visible in a mass spectrum of the main peak either, because the main peak was never the thing that changed.

What happens when the same peptide comes from different sources

The clearest demonstration in the literature predates the current research peptide market by more than fifteen years. De Spiegeleer and colleagues at Ghent University investigated why laboratories could not reproduce published findings on obestatin binding and activation of the GPR39 receptor, including a failure to reproduce reported by the original discoverers. They obtained obestatin from five different manufacturers, the same sources the conflicting research groups had used, and ran identity and impurity profiling by liquid chromatography with photodiode array and fluorescence detection alongside electrospray ionisation mass spectrometry.

One of the products was not obestatin at all. It was a different peptide entirely. Of the remaining products, two thirds were judged unfit for in vitro or in vivo work, defined by the authors as purity below 95 percent or individual impurities exceeding one percent. The authors concluded that these material differences call the divergent published conclusions about obestatin activity into question, and recommended quality control testing before any peptide is used for biomedical research (De Spiegeleer et al., Analytical Biochemistry, 2008, 376(2), 229 to 234).

That is the comparability problem stated in its harshest form. A body of published literature disagreed with itself, and the disagreement traced back to the material rather than to the biology. Every group involved presumably held a certificate saying it had obestatin.

Key Research Findings

• De Spiegeleer et al., Analytical Biochemistry, 2008, 376(2), 229 to 234: obestatin sourced from five manufacturers; one product was a completely different peptide; two thirds of the remainder had purity below 95 percent or individual impurities above 1 percent, rendering them unsuitable for in vitro and in vivo work.
• Hach et al., Pharmaceutical Research, 2024, 41(10), 1991 to 2014: 16 injectable semaglutide, 8 oral semaglutide and 2 injectable liraglutide follow-on products were compared against originator material. Follow-on injectable drug substances and products carried new impurities and new impurity patterns, including high molecular weight proteins, trace metals, anions, counterions and residual solvents.
• In the same analysis, several commercialised follow-on oral semaglutide products contained markedly less semaglutide than the label claim, and dissolution testing showed different release profiles for both the peptide and the SNAC absorption enhancer.
• Neoepitopes were identified in follow-on semaglutide drug substance and drug product, indicating potential immunogenicity differences arising from manufacturing rather than from sequence.
• Fibrillation assays showed increased fibrillation tendency and reduced physical stability in liraglutide follow-on product samples relative to originator product.
• FDA guidance for industry, ANDAs for Certain Highly Purified Synthetic Peptide Drug Products That Refer to Listed Drugs of rDNA Origin (finalised May 2021), covering glucagon, liraglutide, nesiritide, teriparatide and teduglutide: any peptide-related impurity present in both the proposed product and the reference product must be present at the same level or lower, each impurity at or above 0.10 percent of the drug substance must be identified, and a new impurity not present in the reference product must not exceed 0.5 percent of the drug substance.
• FDA regards any polymer composed of 40 or fewer amino acids as a peptide regulated under the Federal Food, Drug and Cosmetic Act rather than a protein regulated under the Public Health Service Act.

The manufacturing process is part of the product

The Hach study is the most detailed public comparison of nominally identical peptides made by different routes. It matters here because the follow-on materials were not counterfeit and were not obviously defective. They were commercial products sold as the same molecule, and an orthogonal analytical battery spanning chromatography with ultraviolet and mass spectrometric detection, inductively coupled plasma optical emission spectroscopy, inductively coupled plasma mass spectrometry, nuclear magnetic resonance, dissolution testing, in silico major histocompatibility complex class II binding prediction and fibrillation assays still separated them from originator material on multiple axes at once.

One caveat belongs in plain view. The study was authored by scientists at Novo Nordisk, which originates both semaglutide and liraglutide and has a commercial interest in the conclusion. That is a real conflict and readers should weigh it. It does not dissolve the findings, because the analytical methods are standard and the observations are specific and falsifiable, but it does mean the work invites independent replication rather than being treated as settled.

The general principle it illustrates is old and uncontroversial in regulated manufacturing: for a molecule of this size and complexity, the process contributes to the identity of the product. Sequence is necessary and not sufficient. Two facilities running different chemistry to the same sequence produce two materials whose minor components differ, and for a peptide those minor components are the part most likely to influence immunogenicity, aggregation behaviour and assay interference.

How regulators handle a change, and why research material sits outside that

In a regulated setting, a manufacturing change triggers a formal comparability exercise. Nothing about that framework reaches research grade material. The ICH comparability guideline written for products subject to manufacturing changes addresses biotechnological and biological products, so a chemically synthesised peptide is governed instead by the drug substance development and manufacture guideline and by change control provisions in good manufacturing practice for active ingredients. Those provisions apply to a licensed manufacturer producing a licensed article. A research chemical supplier is not one, and a reseller who buys finished vials from a contract manufacturer is further removed still.

The FDA generic peptide guidance is instructive precisely because of how demanding it is by comparison. To argue that a synthetic peptide is the same active ingredient as an approved one, an applicant must show impurity-by-impurity that shared impurities are no higher than in the reference product, identify everything at or above 0.10 percent, keep any genuinely new impurity below 0.5 percent, and then characterise and justify even that. It is a nonbinding recommendation rather than a statute, and it applies to five named peptides in an approval pathway, but it establishes what a comparability claim actually costs to support. A supplier asserting that this lot is equivalent to the last one has done none of it.

The gap is structural rather than malicious. Most suppliers do not manufacture. They place orders, and the manufacturer behind those orders can requalify a raw material, change a solvent supplier, or move a step to a different plant without any obligation to tell a downstream reseller, let alone the researcher at the end of the chain. The certificate that arrives with the new lot will look exactly like the previous one because it reports the same tests against the same specification.

What a researcher can actually check

Comparability cannot be established from a percentage, but it can be partly assessed from documents that many suppliers never provide. The most useful artefact is the chromatogram itself rather than the summary line. Two chromatograms from consecutive lots, overlaid, show whether the minor peak pattern held. A shifted retention time on the main peak, a new shoulder, or a redistributed cluster of late-eluting peaks is visible in seconds and invisible in a purity table.

Lot identity is the second checkpoint. A certificate that carries a lot number, a manufacture date and a test date allows a researcher to know whether two vials came from the same production event. A certificate that omits the lot number, or that recycles a single document across multiple purchases, cannot support any comparability claim at all, and that pattern is common. We have covered how such documents fail scrutiny in our work on out-of-specification results and retesting limits.

Third, the specification behind the number needs to be legible. Purity reported against an unstated method is not comparable to purity reported against a different gradient at a different wavelength, which is one reason competent laboratories return different figures on the same vial. The thresholds that determine which impurities even get reported are set out in our discussion of peptide impurity thresholds, and the standards question behind interlaboratory disagreement is treated in our review of peptide reference standards.

Fourth, a change in the impurity fingerprint is not always a process change. It can be carryover from an unrelated compound made on the same equipment, a failure mode examined separately in our analysis of peptide cross-contamination. Distinguishing the two requires mass spectrometric identification of the new species, not a recount of the main peak.

Practical implications for experimental design

Researchers who care about reproducibility already control for reagent lot in antibodies and cell culture serum. Peptides deserve the same discipline and rarely get it. Running a study across two lots without a bridging comparison introduces a variable that will not appear in the methods section and cannot be recovered afterwards. Where a study spans more than one lot, recording lot numbers against experimental blocks at least preserves the ability to test for a lot effect later.

The obestatin case is the argument for this in its complete form. Years of contradictory receptor pharmacology, an inability to reproduce a founding result, and an explanation that turned out to sit in the vial rather than in the assay. The peptides had certificates. The certificates were not comparable, and nobody checked until the literature had already fractured.

Batch specific documentation is the only mechanism that makes any of this checkable, which is why Maple Research Labs publishes third-party certificates tied to identified lots rather than a single generic document. Certificates for tested lots are available on our certificates of analysis page.

Limitations and open questions

The evidence base here is thinner than it should be. The obestatin work examined one peptide across five suppliers in 2008 and has not been systematically repeated across the compounds most common in current research supply. The Hach analysis is recent and detailed but originates with an interested party and covers two GLP-1 analogues rather than the wider catalogue. No published study has yet tracked consecutive lots from a single research peptide supplier over time, which is the specific question this article concerns. Until that work exists, comparability remains an inference drawn from adjacent evidence rather than a measured quantity, and the honest position is that researchers do not currently have a way to verify it from the documents they are given.

For research purposes only. Not for human consumption. Not for diagnostic or therapeutic use.