Problematic images in vendor antibody verification data

This repository hosts documentation of suspicious, apparently manipulated or otherwise problematic antibody verification data in the catalogs of laboratory reagent vendors. As of 4 June 2026, we have documented 462 problematic images across 447 products in the primary antibody catalog of Thermo Fisher Scientific (see the first example spotted by Sholto David on 17 May 2026) and 1 problematic image in the primary antibody catalog of Abcam. Others can submit potentially problematic images that they have spotted in any vendor's catalog at this Google form.

The spreadsheet <date>_catch_all_spreadsheet.csv describes all collected instances of problematic images. The archive <date>_original_images.zip contains all images as originally downloaded from the vendor's website. The archive <date>_annotated_images.zip contains annotations of each image that demonstrate the identified problems. catch_all_spreadsheet.csv has the following columns:

• image_index: a unique index for each image in the spreadsheet
• vendor: the vendor selling the product in question (e.g., "Thermo Fisher Scientific")
• catalog_number: the catalog number of the product in question (e.g., "MA5-12557")
• product_title: the product title as it appears in the vendor's catalog (e.g., "p53 Monoclonal Antibody (DO-7) - MA5-12557")
• product_link: the URL of the product page
• original_image_filename: the filename of the original image as it appears in <date>_original_images.zip
• annotated_image_filename: the filename of the annotated image as it appears in <date>_annotated_images.zip
• summary: a brief description of the identified issues in the image
• identified_by: a description of who originally identified/annotated this image as problematic
• first_identified: the date on which the image was identified/anntotated as problematic
• external_links: URLs to any social media posts/blogs where the observation was original posted online