Grow black mice that are hyperresponsive and white mice that are hyporesponsive (?) while optimizing for minimum genetic distance.

Sell them in equally weighted batches as yin/yang mice ("purely for esthetic reasons, also suitable for lab tests")

For me it looks too easy to have a successful experiment (and article) if you can choose how you poison the animal.

I have always wondered why they did not use a commercial model, and better tested their hypothesis on several, unrelated animal models.