Jocelyn Rice points out perceived shortcomings of the experimental autoimmune encephalomyelitis mouse in modelling multiple sclerosis and in advancing effective human treatments for this disease (Nature 484, S9; 2012; online only). However, her title ('Animal models: Not close enough') seems to cast doubt on the value of animal models in general for developing therapeutic strategies. Even if unintended, this implication undermines efforts to narrow the gap between research and the clinic.

We all share the author's frustration over the lag that separates direct medical benefits from animal research. But to shorten it, we should foster the possibilities of each model, not malign them.

Selective reporting of animal models that fail to deliver anticipated therapies risks promoting misperception among clinical researchers and policy-makers. They are more likely to remember that one mouse model was “worryingly unreliable” for screening multiple sclerosis treatments than they are to recall that mice were crucial in the development of many life-saving therapies.

Examples include ipilimumab, an antibody therapy that extends life in patients with metastatic melanoma; losartan, which reduces aortic disease in patients with Marfan syndrome; and the retinoic acid/arsenic trioxide therapy that saves the lives of patients with acute promyelocytic leukaemia.