Commissioning Your Own Third-Party Test: A Buyer's Procedure for Independent Analysis

Commissioning your own third-party test means you choose the laboratory, you choose the container that is analysed, you decide which questions are asked of it, and the report comes back to you under your own sample identifier. That is the only arrangement in which a result is unambiguously about the material in your possession. This page is a procedure, not a list of warning signs. The many 'vendor red flag' checklists already cover what to be suspicious of; what they skip is what to do once suspicion, or ordinary diligence, calls for evidence you produced yourself. If you are unsure what a supplier's own badge promises, start with what "lab tested" actually claims and come back.

The reason it matters is structural, not a judgement on any supplier. A certificate of analysis is a report on a sample someone else selected, and ISO/IEC 17025 requires every test report to state that its results relate only to the items tested ^[1]. Even a genuine, batch-matched certificate therefore describes the container that went into the instrument, not the one on your bench. Regulated manufacturing has never relied on a supplier's paperwork alone: under 21 CFR Part 211, a receiving party may accept a supplier's report of analysis in place of full testing only if it runs at least one specific identity test itself and checks the supplier's results at intervals ^[3]. Commissioning a test is the research buyer's version of that same control.

When a test is worth commissioning, and when documentation is enough

Independent testing costs money and consumes material, so it should be triggered by a reason rather than by habit. These are the situations where the paperwork alone leaves a question that only an analysis can close.

• A first order from a supplier whose earlier certificates you have never checked against an independent result.
• A certificate that cannot be confirmed with the issuing laboratory, or whose lot string does not match the label.
• Material whose appearance, solubility or behaviour differs from a previous lot of the same item.
• Work whose conclusions will be published, or will rest heavily on the identity or amount of one compound.
• An order large enough that the test is a small fraction of its value.

Documentation is usually enough for a repeat lot from a supplier whose certificates you have already corroborated with your own result, provided the new certificate is batch-matched, dated before dispatch and confirmable with the laboratory that issued it ^[1]. The point of an occasional independent test is partly to earn that trust: one agreeing result makes the next certificate from the same source worth more.

Decide the question before you choose the test

Laboratories sell tests, not answers, so the buyer has to translate a worry into a request. 'Is it real?' is an identity question. 'Is it clean?' is a purity question. 'Is there as much in it as the label says?' is a content question, and it is not answered by a purity figure at all. Each test below answers one of these and is silent on the rest.

Whatever you order, ask the laboratory to name the method on the report. A purity figure is defined by the procedure that produced it — column, gradient, detection wavelength and integration all move it — and the pharmacopoeial chromatography chapter exists precisely because those conditions must be stated and controlled ^[4]. Validation guidance puts the weight on specificity: a method is only useful if it can distinguish the analyte from the impurities likely to be present ^[2].

Identity alone, or a full panel?

Identity plus chromatographic purity is the sensible default. Together they establish that the right molecule is present and that most of what the detector sees is that molecule, which is the question most buyers are actually asking. Identity alone is the cheapest meaningful test and is the right choice when the worry is substitution rather than quality. A fuller panel — adding net content, water and endotoxin — earns its cost when downstream work is quantitative, where the amount of peptide per weighed milligram matters, or cell-based, where endotoxin can produce effects wrongly credited to the compound. Before paying for any of it, it is worth knowing why a bench-top or home kit cannot stand in for a laboratory; the short answer is that none of these questions can be settled without the instruments above.

How much sample a test consumes

Analysis destroys what it measures, so the laboratory's sample requirement is the first thing to ask for, test by test. Chromatographic and mass-spectrometric work needs relatively little material; content determinations and water titration need more; microbiological tests need a container nobody has opened, because opening one outside a controlled environment introduces exactly what the test is looking for. Send sealed, unopened containers wherever possible, since a sealed container removes any question about what happened after it left the supplier.

Before anything leaves your hands, set aside a second sealed container from the same lot as a reserve. Regulated practice keeps reserve samples for the same reason ^[3]: if a result is surprising, the reserve lets you or a second laboratory look again at the same lot without depending on anyone else's material.

Choosing a laboratory, getting a quote

The detail of laboratory selection has its own page in this cluster. The short version: accreditation is only meaningful for the methods listed on the accreditation scope ^[1]; a laboratory that runs peptides routinely is worth more than one that will accept the job; and a report that states measurement uncertainty lets you judge whether two results genuinely differ ^[5]. Ask for a written quote naming each test, the method, the sample requirement, the turnaround from receipt, the report format and whether a rush option exists.

Prices and turnaround move quickly, so treat any figure as a band rather than a promise. As a rough guide, identity plus purity from a laboratory serving this market is typically priced per sample in the low hundreds of US dollars; each added test adds its own cost; accredited work with stated uncertainty generally costs more and takes longer; and turnaround runs from a few working days to around two weeks from receipt, plus transit. The cost article in this cluster sets out the drivers and the lab-stated figures as of this month.

Sending the sample

A result is only as good as the record that links it to your container. Chain of custody is the documented account that a sample stayed under identified control from collection to disposal ^[6], and laboratories accredited to ISO/IEC 17025 are required to identify and protect items from receipt onward ^[1]. The buyer's part is short.

1. Photograph the container label and seal before packing, with the lot string legible.
2. Assign your own sample identifier and write it on the container and on the laboratory's submission form.
3. Seal the outer package with tamper-evident tape and record the date and who packed it ^[6].
4. Declare the contents accurately on the consignment, exactly as the laboratory and carrier require.
5. Keep the reserve container sealed, labelled and stored as the original was.

Reading your result against the supplier's certificate

Compare methods before comparing numbers. Two competent laboratories analysing the same lot will rarely report the same purity figure, because they rarely use the same column, gradient or wavelength; before treating a gap as meaningful, check the chromatographic settings that move a purity figure on both reports. Identity is different: a mass that does not match the stated sequence is a finding on its own. Where uncertainty is stated, a difference smaller than the combined uncertainty of the two results is not a disagreement at all ^[5].

1. Confirm both reports name the same lot, and that your sample identifier appears on yours.
2. Compare identity first: does your mass-spectrometry result match the stated sequence?
3. For purity, note method, wavelength and integration on each report before comparing figures ^[4].
4. Treat a gap of a point or two between different methods as expected; investigate a large one.
5. Record both results, both methods and your conclusion against the lot.

What your own test still cannot tell you

An independent result is about one container, on one day, by one method. It says nothing directly about other containers in the lot, about how the material will behave after months of storage, or about any attribute you did not ask for. ISO/IEC 17025 builds that boundary into the report itself ^[1], and it applies to your commissioned report exactly as it applies to the supplier's. What your own test adds is control over the three things a supplier's certificate cannot give you: which container, which questions, and who received the answer.