The "in mice" caveat is best used for medical treatments. Treatments are less likely to generalize than discoveries. A treatment has to work in every specific.

The "in mice" caveat is best used for medical treatments. Treatments are less likely to generalize than discoveries. A treatment has to work in every specific. There are many, many ways a treatment that works in mice can fail in humans. Any one of them scuppers it. But when you're talking about a discovery, it's probable that it applies to both. It matters less if the human version is somewhat different from the mouse model, as long as it's homologous.