Jessica Bolker's emphasis on choosing the right animal model (Nature 491, 31; 2012) should not undermine the validity of the mouse as a model for human disease. Contrary to her implication, mouse researchers do take genetic background and environment into consideration.

Mouse models used to test therapies should reproduce closely the human disease being investigated — including the response to genetic and environmental factors. But this need not apply to models for understanding disease mechanisms: Nobel-prizewinning discoveries of embryonic stem cells, the mouse major histocompatibility locus and monoclonal antibodies, for example, all relied on experimental mice and not disease models.

Researchers do not use only inbred strains of mice: they use genetically characterized mouse populations and deploy different genetic backgrounds to identify complex inherited traits. There are large collections of recombinant inbred mouse strains, as well as a public database of more than 3,000 traits (E. J. Chesler Nature Neurosci. 7, 485–486; 2004). The Collaborative Cross reference panel, representing eight inbred strains of lab and wild-derived mice, contains twice the genetic diversity of the entire human population and allows high-resolution analysis of phenotypic variations (Genetics 190, 389–401; 2012).

The full potential of the mouse as a model system has yet to be realized. The study of genetic reference populations, along with a library of mutants for every mouse gene, will continue to transform our understanding of human disease.