Poorly characterized antibodies give rise to irreproducible results (see, for example, Nature 527, 545–551; 2015), but so can the use of properly validated antibodies in a non-validated context.

At Aviva Systems Biology in California, we used our highly specific commercial antibodies in western immunoblot assays to test more than 1,000 protein samples provided by the research community. We found that the preparation quality of more than half of these samples failed to meet the technical requirements for a reliable assay signal.

Simple technical factors confounded the electrophoretic resolution or antibody detectability of the researchers' protein solutions. These included unsuitable sample concentrations, buffer incompatibility and the absence of calibration markers or treatment controls. Until uniform western-blotting standards are widely adopted (see J. E. Gilda et al. PLoS ONE 10, e0135392; 2015), there is a risk that data irreproducibility will continue to be the norm.

Antibody-production companies should not be treated as casinos for boosting a researcher's chances of a positive result from such shot-in-the-dark samples.