Two Laboratories, One Peptide, a 9.7% Gap: Our Cross-Laboratory Content Data

Two laboratories measured the peptide content of our MOTS-C 10 mg material and disagreed by 9.7%: Janoshik Analytical reported 11.82 mg in April 2026, and SideChain Analytics of Mississauga reported 12.97 mg on 29 September 2026 under report COA-2026-SC-02984. We cannot say which figure is correct. This page sets out both certificates, three further lots where two laboratories disagreed on our own material, and what we think the plausible causes are.

What did the two laboratories report on the MOTS-C material?

SideChain Analytics, report COA-2026-SC-02984, analysed 29 September 2026. Sample MOTS-207_D, run by HPLC-MS/MS on an Agilent 1290 Infinity II with a 6495 triple quadrupole. Identity confirmed as MOTS-C against a reference standard, with the MRM transition 725.5 › 175.1 as quantifier and retention time in agreement. Chromatographic purity 99.5% on the principal-component peak-area convention of European Pharmacopoeia 2.2.46. Net peptide content 12.97 mg against a 10 mg label, or 129.7%. The laboratory states on the certificate that chromatographic purity bounds peptide-related impurities only — not excipients, counterions or residual water — and that the report carries no pass/fail determination. We commissioned this test.

Janoshik Analytical, report #147077, April 2026. HPLC purity assay with quantitative content. Purity 99.025%, net peptide content 11.82 mg against the same 10 mg label, or 118.2%. This test was commissioned by the manufacturer, not by us.

A separate endotoxin certificate exists and is not in dispute. SideChain report COA-2026-SC-02984-EN, sample ENDO-MOTS-207_D, an 11.9 mg aliquot analysed 23 September 2026 by kinetic chromogenic LAL/TAL assay under USP <85>, returned a result below 0.11 EU/mL against a 1 EU/mL limit set by the test-kit manufacturer. "Below" means under the assay's quantifiable range for that sample, not zero, and the figure is per millilitre of the solution the laboratory prepared. Only one laboratory ran it, so there is nothing to compare it against.

One defect on the SideChain certificates that we have not resolved. The lot-number field on both the purity and the endotoxin report reads reta30-0626-02, a code belonging to a different product, while the sample photograph on each shows a MOTS-C vial labelled MOTSC10-0626-02. We have raised it with the laboratory. Until it is corrected we attach no lot number to these results, and we note here that it weakens the claim that the two laboratories tested the same lot rather than two lots of the same compound. Full detail is on the MOTS-C product page.

How large is a 9.7% disagreement, and is it typical?

It is the same order of magnitude as the only comparable published investigation we know of. In "Testing the Peptide Testers," published 3 September 2025, Rory at graymarket arranged for four vials from a single retatrutide batch to go to different laboratories. Janoshik and Chromate each returned roughly 30 mg; Krause Analytical returned 26 mg on each of two vials. The author describes the gap as "a full 15% difference."

Two things are worth taking from that piece. The first is the magnitude: an inter-laboratory spread in the low to mid teens of per cent on quantitative peptide content is not an anomaly, it is what happens when the same material goes to different places. The second is the author's own framing of the exercise — a small, informal experiment run with their own stock, not a controlled protocol. It does not establish a systematic pattern and does not claim to. Nor does what follows here.

What does not exist, as far as we can find, is a blinded multi-laboratory round-robin on research peptides: identical aliquots of a homogenised batch, coded, distributed to several laboratories, with the results analysed together. That is the standard instrument for separating method bias from material variation, it is routine in other analytical fields, and nobody in this one appears to have run it. The two datasets on this page are one more data point toward that gap, not a substitute for closing it.

Is MOTS-C the only lot where two laboratories disagreed?

No. Four of our lots have been measured by two laboratories, and all four disagree. Three of the four gaps are quantified; the fourth is qualitative because the earlier figure was never published numerically.

Material	First laboratory	Second laboratory	Gap (high ÷ low)
MOTS-C 10 mg lot number disputed	Janoshik #147077, Apr 2026 11.82 mg · 99.025%	SideChain COA-2026-SC-02984, 29 Sep 2026 12.97 mg · 99.5%	9.7% second higher
Tesamorelin 10 mg supplier lot CS-te10-0616	Janoshik #198440, 1 Jul 2026 11.56 mg · 99.174%	Testides ARUA-TESA-10-092026, 28 Sep 2026 9.94 mg · 95.19%	16.3% second lower
CJC-1295 (no DAC) + Ipamorelin lot cjc10-0626-02	Janoshik #209639, Jul 2026 11.03 mg total CJC 5.07 · Ipa 5.96	Testides ARUA-CJCN-10-092026, 2 Oct 2026 10.08 mg total CJC 4.74 · Ipa 5.34	9.4% second lower
GHK-Cu 50 mg lot GHK50-0626-02	Janoshik #210586 "close to label weight" no figure published by us	Testides ARUA-GHK-50-092026, 28 Sep 2026 68.67 mg · 99.93%	not quantified

Gap is the higher figure divided by the lower, expressed as a percentage. "Second lower" means the later certificate, which in three of four cases is the one we commissioned, returned the smaller number. Per-component figures for the blend are shown because a two-API vial has two product peaks.

The direction is not consistent, and that is the most informative thing in the table. On the two lots where Testides and Janoshik both ran, Testides read lower — by 16.3% on Tesamorelin and 9.4% on the CJC-1295/Ipamorelin blend. On the one lot where SideChain and Janoshik both ran, SideChain read higher, by 9.7%. Three pairs is far too few to call a laboratory bias, but it is enough to say that "one laboratory always reads high" does not describe what we have. Whatever is producing these gaps is not a single constant offset.

The purity figures diverge too, and less than the content figures do. MOTS-C moved from 99.025% to 99.5%, half a point. Tesamorelin moved from 99.174% to 95.19%, four points — a much larger relative change than most of these laboratories' quoted precision would suggest, and on the lot with the largest content gap. The CJC blend is not comparable on purity, because the Testides figure of 95.80% is a combined named-component share rather than a single main-peak purity.

What could cause two laboratories to differ by 10% on the same material?

Six candidates are worth taking seriously. We cannot rank them with the evidence we have, and we are not going to pretend otherwise.

1. Different reference standards. A quantitative content figure is a comparison against a standard of assumed concentration. If two laboratories build their curves from different material — different supplier, different salt form, different water content, different assigned potency — the scale itself differs, and every sample read against it inherits that offset. This is the single most likely source of a persistent percentage-level gap, and neither certificate states the provenance or assigned potency of its standard.

2. Different detection principles. HPLC-UV at 214 nm quantifies by absorbance of the peptide bond and aromatic residues; HPLC-MS/MS quantifies by the ion current of a selected transition. These respond to different molecular properties, calibrate differently and have different linear ranges. The MOTS-C pair is exactly this comparison: a UV-based Janoshik assay against a triple-quadrupole MRM method at SideChain. The Tesamorelin and CJC pairs, however, are UV against UV, which weakens any explanation resting on detection principle alone.

3. Counterion and water handling. A synthetic peptide is supplied as a TFA or acetate salt with residual water, and counterion plus water can be a double-digit fraction of the powder. Whether a laboratory corrects for salt and water, and how, directly scales its milligram figure. Two laboratories that both report "net peptide content" may be applying different corrections, or one may be applying none.

4. Sample heterogeneity. These are two different vials, filled at different points in a run, tested five months apart in the MOTS-C case. Fill variation within a batch is real, and our own catalogue shows how wide fill can range. The lot-number error on the SideChain certificate makes this harder to rule out, not easier: we cannot currently demonstrate from the paperwork alone that both laboratories received material from the same fill.

5. Sample preparation and recovery. Dissolution of the cake, dilution accuracy, adsorption of peptide to plastic or glass surfaces, and incomplete recovery from the vial all shift the result, and all of them are laboratory-specific. Peptide loss to surfaces at low concentration is a well-known effect and acts in one direction only — downward.

6. Measurement uncertainty, unstated. None of these certificates publishes an uncertainty budget, and most rest on a single injection. A point estimate without an interval cannot be said to disagree with another point estimate in any formal sense. Part of the apparent gap may be nothing more than the ordinary precision of the method, unreported.

What is a blind submission, and why is it harder to game?

Two of our retatrutide lots were submitted to Janoshik Analytical as unidentified samples on a common GLP-1 peptide panel, and the laboratory had to determine what they were before assaying them. Lot CS-re20-0710, report #207689, dated 30 July 2026: identified as retatrutide, 99.643% purity, 24.51 mg against a 20 mg label. Lot CS-re30-0718, report #210575, dated 5 August 2026: identified as retatrutide, 99.751% purity, 34.20 mg against a 30 mg label. The panel spanned semaglutide, tirzepatide and retatrutide, with no indication on the submission of which of the three each sample was supposed to be.

In a standard third-party assay the identity claim originates with the vendor. A submission form names the compound, the laboratory selects or develops a method for that compound, and the report states the area-percent of the expected peak. The report is honest and the laboratory has done nothing wrong — but it answers one question, "how much of this is the thing you said it was," and takes the premise on trust. A blind submission removes the premise. The laboratory resolves identity first and assays second, so the conclusion "retatrutide" is generated evidence rather than confirmed input.

The failure mode this addresses is specific. The realistic adulteration risk in the GLP-1 family is not an obviously wrong powder — it is a cheaper, closely related peptide from the same family sold as the expensive one. An assay designed around an asserted identity is poorly placed to catch that, because it was built to look for what it was told to find. A panel that forces discrimination between semaglutide, tirzepatide and retatrutide is built to catch exactly it.

What it does not do. Blinding does not make the purity figure larger, more precise or more reproducible. It says nothing about content accuracy, which is the subject of the rest of this page, and these two lots are no more immune to inter-laboratory disagreement than any other — neither has a second laboratory's figure to compare against, because Testides confirmation on both is still pending. Blinding addresses identity, and only identity. It is also not the same as a blinded round-robin: one vendor sending unlabelled samples to one laboratory is a weaker design than a coordinated study across several, and we are not claiming otherwise. Our third retatrutide lot, the 10 mg, was not submitted this way and we do not describe it as blind tested.

What have we not resolved?

Which MOTS-C content figure is closer to the truth. Open. We have no characterised reference material and no replicate data, and we have not sent a third aliquot anywhere.

Whether the two MOTS-C certificates describe the same fill. Open, and made worse by the lot-number field on both SideChain reports naming a different product. Raised with the laboratory; no correction received.

The basis of the GHK-Cu content figure. Open. GHK-Cu is 401.91 g/mol and free GHK is 340.38 g/mol, an 18.1% difference, and a figure computed on one basis and read against the other is wrong by that much. We have asked Testides and have not had an answer, which is why that pair is the unquantified row in the table.

Why the Tesamorelin purity figures differ by four points. Open. Oxidation of the methionine at position 27 between the two tests is a candidate, given three months of storage and transport between them, but HPLC-UV cannot identify the impurity peaks and neither laboratory ran mass spectrometry on this lot. We offer it as a hypothesis, not a finding.

Whether any of this generalises. It does not, on this evidence. Four pairs, three quantified, three laboratories, one vendor's supply chain, no replicates, no blinding of the comparison itself, and no uncertainty budgets. Neither Testides nor Janoshik is accredited, and we claim no accreditation for anything on this page. What would settle it is a coded round-robin on homogenised material across several laboratories, which is not something a single vendor can credibly run on its own stock.

Frequently asked questions

Which MOTS-C number does the product page show?

The SideChain figures, 99.5% purity and 12.97 mg content, are shown as the headline because that is the test we commissioned ourselves and the certificate is verifiable at the laboratory. The earlier Janoshik figures of 99.025% and 11.82 mg remain visible on the same page for comparison. Showing the higher number as the headline is a choice that happens to flatter us, which is why the lower one stays on the page.

Does a 9.7% disagreement mean one laboratory made an error?

Not necessarily. Two laboratories can each run a correct method, correctly, and return figures that differ by this much if their reference standards, counterion corrections, sample preparation or detection principles differ. Neither certificate publishes an uncertainty budget, and a point estimate without an interval cannot formally be said to contradict another. An error is one explanation among several, and we have no evidence identifying which applies.

Is HPLC-MS/MS more accurate than HPLC-UV for peptide content?

More specific, which is not the same as more accurate for quantity. Mass spectrometry with a selected transition distinguishes the target from co-eluting material far better than absorbance at 214 nm, so it is the stronger identity tool. For quantity, both methods are only as accurate as the reference standard the curve was built from, and MS responses can vary with matrix and ionisation efficiency in ways UV absorbance does not. Our Tesamorelin and CJC pairs were UV against UV and still disagreed by 16.3% and 9.4%, so detection principle is clearly not the whole story.

What is a blind submission, in one sentence?

A sample sent to a laboratory without being told what compound it is, on a panel of candidates, so the laboratory has to establish identity from the chromatogram and the mass rather than confirming what the submission form asserted. Our Retatrutide 20 mg and 30 mg lots were submitted this way, under Janoshik reports #207689 and #210575.

Has anyone run a proper multi-laboratory comparison on research peptides?

We have not been able to find a blinded round-robin — identical coded aliquots of a homogenised batch distributed to several laboratories and analysed together. The closest published work is graymarket's "Testing the Peptide Testers" of 3 September 2025, which sent four vials from one retatrutide batch to multiple laboratories and found roughly 30 mg at two of them against 26 mg at a third, and which presents itself as an informal experiment rather than a controlled study.

Why does the lot number on the SideChain certificate name a different product?

We do not know. The field on both the purity report and the endotoxin report reads reta30-0626-02, while the sample photograph on each shows a MOTS-C vial labelled MOTSC10-0626-02. The most likely explanation is a clerical carry-over on the laboratory's side, but we cannot demonstrate that, and until it is corrected we attach no lot number to those results and treat the same-lot assumption as unproven.

What should a reader do with a single certificate, given all this?

Read it as one measurement with an unstated uncertainty, not as a property of the material. Check that the lot number on the certificate matches the vial, that the laboratory is named and the report independently verifiable, and that purity, content and identity are reported as separate determinations. Our guide to how to read a peptide COA works through the fields, and where to get peptides tested in Canada covers commissioning your own.

What could cause two laboratories to differ by 10% on the same peptide material?

Six candidates, in no particular order: different reference standards and assigned potencies, different detection principles such as HPLC-UV against HPLC-MS/MS, different counterion and residual-water handling, genuine sample heterogeneity between vials, differences in sample preparation and recovery including peptide adsorption to surfaces, and ordinary measurement uncertainty that none of these certificates states. Each is set out above.

Sources

1. SideChain Analytics, Mississauga, Ontario — report COA-2026-SC-02984 (identity and purity, HPLC-MS/MS, Ph. Eur. 2.2.46, analysed 29 September 2026) and COA-2026-SC-02984-EN (bacterial endotoxin, kinetic chromogenic LAL/TAL, USP <85>, analysed 23 September 2026). Verification record · Certificate PDF
2. Janoshik Analytical reports #147077 (MOTS-C, April 2026), #198440 (tesamorelin, 1 July 2026), #209639 (CJC-1295 no DAC + ipamorelin, July 2026), #207689 (retatrutide 20 mg, 30 July 2026), #210575 (retatrutide 30 mg, 5 August 2026) and #210586 (GHK-Cu). Reports #147077, #198440, #209639 and #210586 were commissioned by the manufacturer.
3. Testides, Toronto — reports ARUA-TESA-10-092026 (tesamorelin, reported 28 September 2026), ARUA-CJCN-10-092026 (CJC-1295 no DAC + ipamorelin, issued 2 October 2026) and ARUA-GHK-50-092026 (GHK-Cu, reported 28 September 2026). Certificates are verified by document upload at testides.com/verify.
4. Rory, "Testing the Peptide Testers," graymarket, 3 September 2025. graymarket.substack.com/p/testing-the-peptide-testers
5. Finnrick, Methodology — vendor-testing inputs and quantity tolerance bands. finnrick.com/methodology
6. Panda Peptide, Certificates of Analysis (Janoshik verify table). pandapeptide.ca/certificates-of-analysis/
7. Health Canada, public advisory, 9 April 2026: "Think twice before injecting peptides bought online: unauthorized products can seriously harm." recalls-rappels.canada.ca
8. Health Canada, "Health Canada obtains permanent injunction preventing Canlab Research's illegal sale of injectable peptides in Canada," July 2026. canada.ca
9. European Pharmacopoeia general chapter 2.2.46 (chromatographic separation techniques) and United States Pharmacopeia general chapter <85> (bacterial endotoxins test), as cited on the SideChain certificates.

Related reading

Fill accuracy across 17 lots · How to read a peptide COA · Where to get research peptides tested in Canada · MOTS-C research guide · All lot results

Published 19 October 2026 by the thepeptide.ca research team. Certificate data current at that date. If either laboratory issues a correction we will publish it here and date the change.