>>12403565>>12403535>>12404429There was a big protein folding competition that is held every year. There are the experimental structures that have been determined, and then teams use their techniques to try and get as many proteins perfect as possible.
What happened:
we went from a ~60% accuracy to a 92% accuracy. Experimentally determined structures are at about 90% accuracy.
The caveat:
This is based on homology type modeling so it isn't great at edge cases and things like IDP's (I think). But if we can get google to unhinge it's asshole to let others use it we could be in the cusp of the true bio-revolution to be honest. Designer proteins have been getting easier and easier, now we are having our computational methods actually overtaking conventional experiments
