The reproducibility crisis in life sciences has been pinned on many causes; poorly validated antibodies are the largest single contributor. The International Working Group for Antibody Validation (IWGAV) has proposed a five-pillar framework that’s now adopted by major journals and funders.
Why antibody validation is so often skipped
Vendors validate antibodies for “applications” (WB, IHC, IF, FC) but rarely for specific cell types, fixation conditions, or epitopes within isoforms. Many “validated” antibodies cross-react, recognize wrong proteins after fixation, or simply don’t work for your sample type. Doing your own validation is the only way to be sure.
The five pillars of antibody validation
1. Genetic strategy (the gold standard)
Compare signal between wild-type cells and a CRISPR knockout (or shRNA knockdown) of the target gene. The signal should disappear in the knockout. This is the most rigorous test because it directly addresses specificity.
2. Orthogonal strategy
Compare antibody-based detection to an independent method — typically mass spectrometry or RNA expression. Concordance across orthogonal methods supports specificity.
3. Independent antibody strategy
Use two or more antibodies recognizing different epitopes of the same target. Concordant results across antibodies argue against epitope-specific artifacts.
4. Tagged protein expression
Express the target with an epitope tag (FLAG, HA, V5) and compare anti-tag detection to anti-target detection. Co-localization or co-migration supports specificity.
5. Immunoprecipitation followed by mass spectrometry (IP-MS)
Pull down with the antibody, identify proteins by MS. The intended target should be the dominant hit.
Application-specific validation
Validation must be done for the application you’ll use. An antibody validated for western blot may not work for IHC, and vice versa.
- Western blot: Knockout/knockdown comparison; correct molecular weight
- IHC/IF: Knockout tissue or peptide blocking; expected localization
- Flow cytometry: Isotype control; correlation with mRNA in known-positive populations
- ELISA: Standard curve linearity; spike-in recovery; cross-reactivity panel
- Immunoprecipitation: IP-MS or IP from knockout cells
Practical workflow
- Check vendor data critically — what cells, what application, what controls?
- Search Antibodypedia, CiteAb, and recent literature for independent validation
- Order 2–3 antibodies from different vendors targeting different epitopes
- Test side-by-side with positive and negative controls
- Document everything — lot number, dilution, blocking buffer, exposure conditions
Red flags in published data
- Bands at unexpected molecular weights without explanation
- No loading control or only a poorly resolved housekeeping protein
- No knockout/knockdown comparison
- Lot number not reported
- Antibody used for multiple applications without app-specific validation
Recombinant antibodies as a long-term solution
Recombinant antibodies are produced from cloned heavy and light chain genes, eliminating lot-to-lot variation. They cost more upfront but produce more reproducible results across years and labs.
Validating antibodies in your specific system before committing to a major experiment is the cheapest insurance you’ll buy. The five pillars give you a structured way to ask: do I really know what this antibody is detecting?
