And a condition we can induce in another animal. If we can't cause it in another animal then we can't test a drug against it. There are a lot of diseases that can make no progress under this model which we can not induce in animals or humans.

Fascinating! Reminds me how machine learning as a field is often focused on tasks that can be quantitatively evaluated using large datasets. There are a lot of tasks that do not fit this model, especially interactive tasks.

This isn’t a medication it’s a mechanism study.

I would expect not many and certainly far less than the number of medications that worked in animals but did not work or harmed humans.

Aspirin is famous for having teratogenic effects on rabbits and would never have been approved for humans if it were to be tested nowadays

Why? Plenty of teratogenic drugs are approved and in routine use in non-pregnant patients.

Any recent example?

Not really sure what you're getting at, teratogenicity only matters if a person is pregnant and lack of teratogenicity is in no way a requirement for FDA approval.

There are many commonly used drugs with teratogenic risks that we discontinue in pregnancy or counsel the patient on, picking randomly from a list of medications that were approved in 2023: sparsentan

I'm not sure if I'm missing something but why are you thinking teratogenicity matters when the vast majority of patients are not pregnant at a given moment?