We've been taught about this in bioinformatics 20 years ago. That's the eventual goal - to simulate human biochemistry close enough, so you can test novel drugs without using actual humans or animals. We're getting closer and closer to that every year, but yes, it's a very hard problem to solve.
