That's not really where most biotechs die.

You don't move a drug into clinical trials unless you can replicate the initial studies, which is almost always "drug x cures y in animal z."

They die when they try to determine if "drug x cures y in humans."