It seems like the most elementary of research principles: Make sure the cells and reagents in your experiment are what they claim to be and behave as expected. But when it comes to antibodies—the immune proteins used in all kinds of experiments to tag a molecule of interest in a sample—that validation process is not straightforward. Research antibodies from commercial vendors are often screened and optimized for narrow experimental conditions, which means they may not work as advertised for many scientists. Indeed, problems with antibodies are thought to have led many drug developers astray and generated a host of misleading or irreproducible scientific results.
This week, more than 100 researchers, antibody manufacturers, journal editors, and funders met in Pacific Grove, California, to hash out standardized approaches to antibody testing. “Cell authentication is a walk in the park compared to what we need to do with antibodies,” says Leonard Freedman, president of the Global Biological Standards Institute (GBSI), a Washington, D.C.–based nonprofit that advocates for better basic research practices and that sponsored the meeting. In the coming months, the attendees hope to come up with a scoring system that will identify the most reliable antibodies for a given type of experiment and ultimately (they hope) make results more reproducible across experiments.
Antibodies are typically made in animals such as rabbits or goats, by injecting a protein of interest and waiting for the animal’s B cells to respond to the foreign molecule with the Y-shaped proteins, which can be isolated from its blood. But batches of the same antibody from different animals may cross-react with different proteins. And it’s hard to trace a given batch to its origin, because antibodies are often relabeled and resold by another vendor under a new name, Freedman says.
Validating the roughly 2.5 million commercially available antibodies that react to human proteins is a mammoth task, acknowledged Mathias Uhlén, a microbiologist at the Royal Institute of Technology in Stockholm and a member of the meeting’s steering committee, in a press briefing. He expects that antibody providers would do the testing and publicize their scores to make their products more competitive. Journals and funding agencies would in turn need to favor research that uses well validated antibodies, he said.
This week’s meeting is one of several efforts to attack what some consider a reproducibility crisis on the antibody front. In an online survey of 504 researchers published by GBSI this summer, more than half of respondents said they had received no training on how to validate antibodies. Experts in an ad hoc international antibody validation working group published a commentary earlier this month in Nature Methods advocating for application-specific antibody testing. Working groups from the meeting intend to publish white papers on their scoring system in journals over the next 6 months.
With reporting by Meredith Wadman.