Well, if you mean 3D model in the computer, I make them a lot and I think they’re all pretty good. We judge how good they are by the “hit rate”. That means we count how many compounds I’ve predicted to be good are actually good. Many of the published models have hit rates in the high single digits or low double digits. Mine are almost always in the double digits, and not always very low….. My personal best was a 45% hit rate.
If you mean 3D models that we can actually test in a lab rather than in a computer, we have several really good models where we grow skin cells in a petri dish to simulate real skin. Then, we add compounds to them and watch what happens.
Comments