I hate to say this, but protein folding projects aren't good for any practical purposes (or show me a breakthrough). It is good for PR (public participation, awareness of the topic), but like SETI (FFT on essentially CBR noise from space) it is quite futile: while in theory it could work, it is really not a reasonable way to use scarce resources.

It is literally burning electricity (oil, coal, Fukushima) for no real reason (MS wants to patent computational heating?)

We have no clue how proteins work/interact (this is why epigenetics is hyped these days). We have some "educated guess", but it is mostly data harvesting from public databases and then doing some simple Bayes or correlation analysis, without much scrutiny on the harvested data itself (ie. was it made up using multiple imputation?)