Third-party testing versus internal QA, and why we publish the COA before it lists.

Most peptide suppliers, including most of the European peptide suppliers we compete with, run their own quality assurance. The peptide arrives from the synthesis lab, the supplier tests it, the supplier issues a certificate of analysis on the supplier's own letterhead, the peptide ships.

This is not wrong. Internal QA can be done well, and at pharmaceutical manufacturers governed by GMP it is the legally required model. But for a research compound supplier sourcing peptides from external synthesis labs, internal QA has a structural problem: the supplier has financial incentive to find the result they want to find. Even an honest operator running careful methods is one informal interpretive decision away from publishing a number that is more flattering than the underlying data supports.

Third-party testing solves this by removing the integrity question from the supplier's hands.

The mechanics

Cresten works with Janoshik Analytical s.r.o. in Brno, Czech Republic. Janoshik is an ISO/IEC 17025 accredited testing laboratory whose accreditation status is recorded in the public register of the Czech Accreditation Institute (Český institut pro akreditaci, CAI), and which has a long history serving the European peptide research community. They are accredited of Cresten in the operational sense (no shared ownership, no shared management, no commercial dependency, no contract guarantees). They charge us per-batch testing fees the same way they charge any other client.

The process for each batch:

1. Bulk peptide arrives at Cresten from the synthesis lab.
2. A representative sample is drawn under controlled sampling protocol and shipped to Janoshik.
3. Janoshik runs the standard four-panel analytical battery: HPLC purity, LC-MS identity confirmation, endotoxin LAL, and microbial bioburden by enumeration. The methods are documented and the integration parameters are reported. Where the buyer requires extended analytical scope (residual solvents, heavy metals, or related-substance profiling beyond the standard panel), additional testing is commissioned per request.
4. Janoshik issues a certificate on Janoshik letterhead, signed by the analyst, dated, with batch traceability.
5. The certificate is published on the public website before it lists for sale.
6. The batch is then released for vialling, lyophilisation, and packaging.

The key procedural feature is the publish-before-listing step. By the time a customer can place an order for a batch, the certificate for that batch is already publicly visible and downloadable. The supplier has no opportunity to hold back an unfavourable result, no opportunity to retest until the result is acceptable, no opportunity to release product first and reconcile certification later.

What this costs

Janoshik charges approximately 180 to 250 euros per batch for the standard panel, depending on the specific compounds and assay requirements. For a 100-vial production run, that is roughly 2 euros per vial added to cost of goods. The cost is built into Cresten pricing and is not optional.

Internal QA at the same depth would cost less per batch in marginal terms (existing instrument time, existing analyst hours), but it would not deliver the same trust signal. The 2 euros per vial is what trust costs.

What this does not solve

Third-party testing is necessary, not sufficient. It tells you what is in the vial. It does not tell you anything about how the vial was handled, how it was packaged, how it was shipped, or how it was stored. The standard supplementary discipline at Cresten covers cold-chain shipping for compounds that require it, lyophilisation under controlled atmosphere conditions, sealing under nitrogen for compounds vulnerable to oxidation, and stability indication for compounds tested at the time of release rather than only at the time of synthesis.

Accredited testing also does not solve the buyer-side problem of confirming that the vial they received actually came from the batch the certificate covers. The batch number on the vial cap should match the batch number on the certificate, and the QR code on the vial should resolve to the public verification page for that batch on the Cresten site. If the codes do not match, the vial did not come from us, regardless of what the packaging says.

Why publish the COA before listing

Most suppliers who do third-party test still publish the certificate after sale. The customer pays for the vial, receives the vial, opens the packaging, and finds the COA inside. This is the model that pharmaceutical manufacturers use under GMP. It works for medicinal products because the customer is a hospital pharmacy with the standing to refuse unacceptable batches.

For research customers, publishing post-sale creates an asymmetry. The customer commits to purchase based on a label and a description, then learns the actual batch quality after the transaction is complete. Returns are difficult, refunds rare, and the customer's only recourse for a marginal batch is to absorb the loss or run their experiment with worse material than they expected.

Publishing the COA pre-sale removes the asymmetry. The customer sees the actual batch numbers before they decide. If the purity is the reported percent on this batch and the customer wanted 99.5 minimum, they buy. If the purity is 98.4 on this batch and the customer wanted 99.0 minimum, they wait for the next batch. The decision is informed.

This also creates an internal pressure on Cresten that we want to be there. We cannot ship a batch with marginal results because the marginal results are visible before any sales happen. The only batches that get listed are batches we are willing to publish. The publishing discipline is the testing discipline.

The short version

Internal QA is a question of trust between supplier and customer. Third-party testing is a question of evidence. We chose evidence because it scales, because it survives scrutiny, and because the additional 2 euros per vial it costs is the cheapest insurance available against the question we never want to be unable to answer: how do you know.

We know because Janoshik tested it, and Janoshik does not work for us.