How to Verify a Peptide Supplier COA and Raw Data

Start with the identity chain

Compare the purchase order, vial label, packing list, COA, chromatogram, and LC-MS report. Product name, sequence, terminal modifications, salt form, batch number, and test date should agree. Trade names are not enough. CJC-1295 with DAC and without DAC require different structural specifications. TB-500 cannot be accepted as a defined analyte unless the supplier discloses the sequence. A small molecule such as ibutamoren should not carry a peptide sequence field.

Recalculate theoretical mass independently. Check whether the supplier reports average or monoisotopic mass and whether lipid, linker, disulfide, terminal amidation, acetylation, or counterion has been included. A correct mass copied from a catalog does not prove the spectrum belongs to the lot, but an incorrect theoretical mass immediately shows that technical review failed.

Inspect the chromatography as a sequence

Request the full batch run, including blank, system-suitability injections, reference, sample replicates, wash, and the injection after the sample. Standalone PDFs hide carryover and reintegration. File timestamps should follow a believable order. Sample names should not change between acquisition and report without an explanation.

Review the complete acquisition window. Cropping the first two minutes can hide polar deletion sequences; cutting the end can hide hydrophobic impurities or lipidated-peptide carryover. Ask for integration events and manual changes. A baseline adjusted separately for each batch is not a controlled method.

LC-MS evidence needs raw charge states, calibration status, deconvolution range, tolerance, and adduct interpretation. A spectrum containing only a pasted molecular-weight label is not raw data. For modified peptides, ask what part of the structure the method actually confirms. Intact mass does not prove attachment site or stereochemistry.

Check whether the same data were reused

Overlay chromatograms from several lots. Real replicate runs look similar but not pixel-identical. Identical noise, baseline drift, retention time to four decimal places, and impurity pattern strongly suggest the same file was reused. Image metadata and PDF creation dates can support the review, although metadata alone is not decisive.

Search the spectrum title, instrument sample ID, and unusual formatting across older certificates. Resellers often receive one generic report and issue new COAs under their own batch codes. Ask for the mapping between reseller batch and manufacturing batch. If the supplier refuses, treat traceability as absent rather than “confidential.”

Minimum supplier audit package

• Legal manufacturing site, laboratory location, scope of work, and subcontracted testing disclosure.
• Approved specification with version, test methods, acceptance criteria, and reporting basis.
• Raw HPLC and MS data with sequence context, audit-trail availability, and analyst/reviewer approval.
• Peptide content, water, counterion, residual solvent, and fill data separate from area purity.
• Reference-standard qualification and instrument calibration/maintenance evidence.
• Deviation, out-of-specification, change-control, complaint, and data-retention procedures.

A GMP certificate deserves its own verification. Confirm the issuing authority, facility name, address, scope, issue date, and database status. A valid certificate for packaging does not qualify peptide synthesis. ISO accreditation may cover only named test methods. “GMP-compliant” without an auditable scope is marketing language.

Independent testing without wasting the sample

Risk-rank the incoming panel. New suppliers and high-value modified peptides justify stronger orthogonal identity and quantitative content work. A known supplier with stable trend data may move to reduced testing, but identity should not disappear. Send blind or coded samples when practical and keep enough material for investigation.

Tell the external laboratory the exact expected structure without giving it the supplier’s observed answer. Otherwise confirmation can become a search for the expected peak. For blends, provide each component and expected ratio; one total purity number is not sufficient.

FDA data-integrity inspection materials emphasize comparison of reported results with original chromatograms and laboratory records. RUO purchasing is not pharmaceutical release, yet the same principle exposes recycled files and selective reporting.

Do not approve a supplier solely because one independent sample passed. Quality systems predict whether the next batch will be controlled. Peptides Archive can help research buyers build a proportionate document request or review a redacted COA package. The service concerns RUO identity, data integrity, and procurement only; it is not a human-use endorsement.

Put expectations into a quality agreement or purchase specification before the order. Define notification periods for sequence, resin, coupling reagent, purification, analytical method, subcontractor, and manufacturing-site changes. Ask how reprocessed or repurified material is identified. A supplier who changes a gradient after an out-of-specification result should document the investigation and method rationale rather than issue a cleaner replacement trace. Trend retention, purity, content, water, and major named impurities across batches. Slow drift is often visible several lots before a formal failure.

Video calls are useful only when tied to evidence. Ask the laboratory to show the instrument sequence, controlled method, sample-login record, and raw-data location rather than giving a general facility tour. Photographing a chromatograph proves ownership, not that it generated the submitted result. Redacted records can protect confidential details while retaining dates, identifiers, and approval history needed for traceability.

Primary records and verification routes

Use the primary paper, current regulator record, or lot-linked analytical file for the claim it supports. A search result is a route to evidence, not evidence itself.