How PeptideScores verifies a COA
The step-by-step internal process for reviewing each certificate before it reaches a vendor page.
This guide documents what happens inside PeptideScores when a Certificate of Analysis is received. It is distinct from how to read a COA (which teaches you to verify a certificate yourself) and from how PeptideScores uses a COA (which explains how evidence relates to scores). Here we describe the review process itself — what is checked, in what order, and what each check can and cannot establish.
What this process is, and what it is not
Our COA review is document inspection. We examine the document itself: the named laboratory, the tested material, the lot and date, the methods used, the reported results, and whether the document is internally consistent. That is the only thing our review process does.
It is not laboratory authentication. We do not contact the issuing laboratory to confirm that a document is genuine. It is not independent testing. We do not commission laboratory tests of any vendor's products. The steps below describe what we do check; the lab verification guide describes what you can check independently.
The ten-step review process
1. Document intake
When a COA is submitted or discovered, the document is stored in our documentation database with a record that includes the source (vendor-supplied, publicly linked, or independently discovered), the file URL, and a timestamp. Every document is assigned a unique internal ID. At this stage the document has not been reviewed; it is simply catalogued.
2. Label and metadata extraction
The document label is extracted and cleaned. Vendor name prefixes are stripped so that a label like "VendorName_BPC157_5mg_COA" becomes "BPC-157 5mg." Where possible, the following metadata is extracted from the document text: the testing laboratory name, the lot or batch number, the report date, the test types identified, the purity percentage, the labeled strength, and the measured content. Extraction is automated but reviewed: if a field cannot be extracted automatically, it is left blank and flagged for manual entry.
3. Compound tagging
The document is matched to a compound in our compound catalog using label text, known aliases, and vendor inventory cross-referencing. If the label mentions "BPC-157" or any of its synonyms, the document is tagged with that compound ID. Documents that cannot be confidently matched to a known peptide compound are excluded from the COA scoring system and flagged for manual review. Non-peptide documents are excluded entirely.
4. Lot and date verification
The extracted lot number and report date are checked for plausibility. A COA without a lot number is not automatically rejected, but it receives reduced credit in the Batch Integrity subscore because it cannot be traced to a specific production run. A COA without a date is flagged as undated and receives reduced credit in the COA Freshness subscore. Where a date is present but ambiguous (for example, a label that reads "03/04/26"), the format is checked against the laboratory's known dating convention before a date is assigned.
5. Reused-report detection
The document is compared against every other COA on file for the same vendor. This check looks for: identical file hashes (the same PDF uploaded twice), identical lot numbers across documents labeled as different compounds or batches, and identical purity values with identical chromatography data across supposedly separate reports. A reused report is a serious flag: it means the vendor is presenting a single test as evidence for multiple products or batches. When detected, the duplicate document is not scored, and the vendor's Batch Integrity subscore is penalized.
6. Lot mismatch detection
If a vendor lists a product with a specific lot number on its website or in its inventory, and the COA on file references a different lot number, this is a lot mismatch. The COA may be genuine, but it does not document the lot being sold. Lot mismatches are flagged internally. The COA is still scored for its testing content, but the Batch Integrity subscore is reduced because the documentation does not trace to the current product.
7. Missing chromatogram flag
A COA that reports an HPLC purity percentage should include a chromatogram or at minimum a peak table. A report that states "99.2% pure" without any supporting chromatographic data is flagged as missing chromatogram evidence. The purity value is still recorded, but the Verified Purity subscore is reduced because the number cannot be independently verified from the document alone. The same principle applies to LC-MS identity claims without mass spectra.
8. Laboratory cross-referencing
The laboratory named on the document is checked against our internal database of known testing laboratories. If the laboratory is recognized (for example, Janoshik, ILS Labs, or other established independent facilities), the Testing Independence subscore receives full credit. If the laboratory is unknown, affiliated with the vendor, or operates in-house, the subscore is reduced. We do not contact the laboratory to verify document authenticity; that step is described in our guide on verifying lab claims, and it is something readers can do independently.
9. Internal consistency check
The document is reviewed for internal consistency: do the reported test types match the methods listed? Do the results align with the acceptance criteria? Does the conformity assessment match the individual test results? A document that states "conforms" but shows an out-of-specification result for one test is internally inconsistent and is flagged. Internal inconsistency does not automatically mean the document is fraudulent, but it reduces confidence and may delay approval.
10. Approval, pending, and publication
After the review steps above, the document enters one of three states. Approved: the document passes review and is eligible for public display on the vendor page. Pending: the document has been received but review is incomplete or metadata extraction needs manual correction. Pending documents are not shown publicly. Excluded: the document failed one or more checks (reused report, non-peptide, unreadable) and will not be displayed or scored. Approval is an editorial decision made by this site; it is not a statement by the laboratory.
What the process catches
The checks above are designed to catch the most common documentation problems we encounter:
- Reused reports — the same COA presented for multiple products or batches
- Lot mismatches — a COA that documents a different lot than the one being sold
- Missing chromatograms — purity claims without supporting analytical data
- Undated or stale documents — certificates that cannot be tied to a specific testing date
- Internal inconsistency — conformity claims that contradict individual test results
- Non-peptide documents — certificates for compounds outside our scope
What the process does not catch
Document inspection has limits. It cannot detect a professionally forged certificate that is internally consistent, carries a real laboratory's name, and includes plausible results. It cannot confirm that the sample tested was drawn from the vendor's stock rather than prepared for the purpose of testing. And it cannot verify that the lot described in the report is the lot you would receive if you ordered today.
These are not gaps we are complacent about — they are limits of what document inspection can establish. The lab verification guide covers the steps a reader can take to check a certificate beyond what our review does, including verifying laboratory accreditation against public directories.
How review results affect scores
The outcome of the review process feeds directly into the Purity & Testing category of the composite score. The scoring methodology guide describes exactly how each subscore is calculated. In summary: approved documents contribute to COA Access, Testing Independence, Batch Integrity, Verified Purity, Testing Breadth, and COA Freshness. Reused reports, lot mismatches, and missing chromatograms reduce specific subscores rather than invalidating the entire document.
Where to go next
Full scoring methodology
Exact weights, subscore calculations, edge cases, and a worked example.
Batch-specific vs catalog COAs
Why a single undated PDF is weaker evidence than per-lot certificates.
How to verify a lab claim
Check an accreditation claim against a public directory yourself.
Testing panels explained
What HPLC-only, HPLC + MS, and full QC panels each answer and leave open.
Research & Informational Use Only. PeptideScores does not sell, endorse, or facilitate the purchase of any compound for human use. All compounds referenced are intended for laboratory research purposes only. We review vendors; we do not provide dosing guidance, medical advice, or treatment recommendations.