How we vet a lab.

Minimum analytical capability we require

A lab is not considered as a primary partner for Open Assay testing unless it holds, in-house and operated by its own staff, all three of the following:

• Reverse-phase HPLC with at least UV detection at 214 nm (peptide bond absorbance) and preferably also 280 nm (aromatic residues). 214 nm is the standard wavelength for peptide purity determination because nearly every amino-acid backbone absorbs there, while only Trp, Tyr, and Phe absorb meaningfully at 280 nm.¹
• Mass spectrometry capable of identifying intact peptide ions in the relevant m/z range — LC-MS (ESI, typically ion trap or Q-TOF) or MALDI-TOF. For most peptides in our coverage universe (MW 500 - 5000 Da), ESI-TOF with sub-ppm mass accuracy is sufficient to discriminate the labeled peptide from common synthesis side-products.
• Endotoxin testing via Limulus amebocyte lysate (LAL) or recombinant Factor C (rFC). The USP Bacterial Endotoxins Test chapter <85> defines the validated sensitivity thresholds; we require a lab with a demonstrated lower limit of quantitation of at least 0.05 EU/mL on the kinetic-chromogenic or rFC platforms they propose to use for our samples.²

A lab that brokers out any of these three to a subcontractor is acceptable only if the subcontractor is also named, is itself vetted against these criteria, and appears on the public COA as the performing lab for the relevant section. We do not accept opaque subcontracting.

Reference standards

For identity and purity claims to be defensible, a reference standard of the target peptide must exist and must be the same material the lab uses to build its calibration curve. Where USP, EP, or NIST reference materials exist for a peptide we cover, we require the lab to use them. Where they do not — which is common, because many research peptides have no pharmacopoeial monograph — we require the lab to disclose the reference material's source, lot, certified purity, and the method by which that certification was established (typically a certificate of analysis from a characterized supplier plus in-house confirmation by orthogonal methods).

Every published COA from our program will name the reference material lot used for each identity and purity call. A peptide tested against an uncharacterized in-house standard will not be published as a pass; it will be published as reference pending and re-run when a traceable standard is available.

Blinded-pair spiking

Before a lab runs its first live Open Assay sample, we qualify them with a blinded-pair spike test. We prepare two vials of reference-grade peptide at identical concentration, label them with barcodes from our standard sequence, and include them in a batch alongside mock samples we have deliberately spiked with characterized impurities — typically a 1.5 - 2.0% solution of a known synthesis by-product, a 0.5% residual scavenger, or a deamidated analog prepared in-house. The lab is not told which barcodes are neat standard, which are spiked, or what the spike is.

A lab passes qualification when:

• Duplicate neat standards return within-lab purity values within 0.5 percentage points of each other.
• Spiked samples return purity values within 0.5 percentage points of the prepared spike level and resolve the impurity as a distinct peak.
• Observed monoisotopic mass for the intact peptide falls within ±0.5 Da of theoretical [M+H]⁺ for ion-trap instruments, or within 10 ppm for Q-TOF / Orbitrap instruments.
• The lab returns an instrument-level raw file (not only a PDF summary) for every barcode in the qualification batch. We audit the chromatogram and spectrum directly.

A lab that misses on any of these is not rejected outright — we discuss the miss, request a method adjustment where appropriate, and re-run the qualification batch at our expense. A lab that misses on two consecutive qualification batches is not used.

Inter-lab comparison

Even a well-qualified lab can drift. Any time we begin using a new peptide chemistry, any time an instrument is reported as having been serviced or re-columned, and on a rolling six-month cadence regardless of those triggers, we run an inter-lab comparison: one blind-purchased sample is split across two independent qualified labs, and both results are compared. Acceptable agreement is defined in advance per-method:

• HPLC purity: within 1.5 percentage points absolute.
• Mass accuracy: same identity call, with observed masses within each lab's declared instrument tolerance.
• LAL / rFC: same pass / flag category, with values within half an order of magnitude on a log scale when both are above the limit of quantitation.

Disagreement outside these windows triggers an investigation before any result for the affected peptide chemistry is published. The investigation is logged, and its conclusions become part of our public corrections record whether or not a previously-published result is affected.

What we look at beyond the instruments

Capability on paper is necessary but not sufficient. Before signing a services agreement with a contract lab, we also verify:

• Accreditation status — ISO/IEC 17025 for the specific methods we will use, or equivalent (A2LA, ANAB). Scope certificates are reviewed to confirm the methods we commission are within the accredited scope, not adjacent to it.
• Whether the lab accepts a strict no-comment-to-vendor clause. Any lab that would confirm or deny to a vendor whether a given barcode was their sample cannot be used.
• Raw-data access: we require the ability to receive unprocessed instrument files (.d, .raw, .lcd, etc.) and the lab's processing method as a separate artifact. A lab that delivers only a final PDF without raw access is not used.
• Data retention: raw files retained by the lab for at least five years, with a written commitment to honor re-analysis requests from us during that window.

Where this stands today

Our buyer-alias and chain-of-custody infrastructure is operational. The lab vetting pipeline described above is being run against two candidate partners — a peptide-specialty analytical provider and a Boston-area CRO with FDA-registered LAL capability. Until one or both of those vettings closes and a services agreement is executed, Open Assay does not publish its own commissioned testing. Coverage on supplier pages cites publicly-available third-party COA data from Janoshik Analytical, Freedom Diagnostics Testing, and Finnrick Analytics with full attribution. See the methodology overview for the transition plan.

¹ Kuipers & Gruppen, "Prediction of molar extinction coefficients of proteins and peptides using UV absorption," reviewed in peptide purity analysis literature (PMID 9004493).
² United States Pharmacopeia, General Chapter <85> Bacterial Endotoxins Test.