Source Composition Sensitivity in an Archive of Peptide Quantity Reports

Background. Public certificates of analysis can support descriptive research about reported peptide quantities, but a large archive may be dominated by one source. Numerical consistency does not establish representative sampling or accurate extraction from every original report. Methods. We replayed a frozen October 3, 2026 extraction covering 12 exact compound categories. After date, duplicate, source and quantity eligibility rules, 29,604 report records remained. The primary outcome was a recorded measured quantity more than 10% below the extracted nominal amount. We compared the pooled archive with the same archive excluding its dominant source, varied the numerical threshold, aligned the latest eligible test date, and examined alternative duplicate rules. These were descriptive analyses of an already inspected dataset, without preregistration or newly performed laboratory assays. Results. The dominant acquisition source contributed 26,445 records, or 89.3%. The primary outcome occurred in 860/29,604 records (2.91%); excluding that source left 74/3,159 (2.34%). Source omission changed individual compound comparisons: GHK-Cu increased from 113/2,546 (4.44%) to 29/260 (11.15%). The pooled proportions were 7.78%, 2.91% and 0.79% at thresholds of more than 5%, 10% and 20% below nominal amount; corresponding proportions after source omission were 5.29%, 2.34% and 0.89%. A common date cutoff retained the overall difference. Alternative duplicate rules yielded a pooled primary-outcome proportion of approximately 2.9%. Conclusions. Source selection materially changes some descriptive quantity comparisons, while its overall direction depends on the numerical threshold. These results concern an archive of extracted report records. They do not estimate product failure probabilities, establish source-specific laboratory performance or validate the quality of products sold. Public aggregate materials permit figure replication; the proprietary source corpus and extraction pipeline remain unavailable for independent reproduction.