Steve Perrin's call to make mouse studies work (Nature 507, 423–425; 2014) resonates with the goals and practices of the International Mouse Phenotyping Consortium (see www.impc.org).

The consortium's ten-year goal is to generate a 'knockout' mutant for every gene in the mouse genome in an effort to characterize the phenotype that each gene confers (S. D. Brown and M. W. Moore Mamm. Genome 23, 632–640; 2012). Standardized phenotyping protocols (www.impc.org/impress) have been carefully designed and validated by the consortium to provide robust, reproducible information, and a statistics advisory group reviews the data and procedures.

All mutant lines and phenotype data are freely available, including summary data for a cohort compared with multiple controls. To overcome any potential issue of publication bias, we include all negative results as well as positive ones.

Careful evaluation of mouse models and their standardized phenotypes, as well as confidence in the reproducibility and validity of summary and individual phenotypic data, is critical to fostering successful preclinical studies.