Wow, this really IS a great community here. I would never have expected to see work-related stuff on a multi-boxing forum. As the OP stated, being able to predict protein structures is the key to designing novel drugs. Unfortunately, it is extremely complex so if you can spare some GHz-hrs, give it a shot. See
http://en.wikipedia.org/wiki/Levinthal_paradox
if you're interested.
Cheers!