Thanks for sharing this. I thought many a times that the huge computational power used in mining may be used for genetic breakthrough.
The problem is that you need to being able to simply prove that a specific piece of work that has been performed and submitted by a miner has actually been done according the "rules". With Bitcoin this is easy "Find an input into hash-function X whose hash-result is lower than [currentTarget]" Done. Work submitted in this way can be verified in milliseconds. How do you do that with research of protein-folding, for example? How do you know the submitted results are correct?