David Baker‘s group has had another success for FoldIt, their computer game for folding proteins. As reported in Nature News, Victory for crowdsourced biomolecule design, FoldIt players managed to improve the computer-designed enzyme that catalyzes Diels-Alder reactions. I’ve not read the paper itself (Increased Diels-Alderase activity through backbone remodeling guided by FoldIt players), which is hidden behind a paywall (though UC has a library subscription, so I do have access).
The initial impressive feat from Baker’s group was that they designed an enzyme de novo using Rosetta design (this was reported in 2010). The FoldIt addition was an 18-fold increase in activity, which came from targeted redesign (redesigning a loop to increase contact between the protein and one of the reactants, stabilizing the new loop, and so forth). The FoldIt players were not not choosing the goals of the redesign (that was done by postdocs in Baker’s group), but they were exploring the protein space in ways that conventional optimization programs do not (such as adding 13 amino acids).
This is not the first success for FoldIt. They had a Nature paper in 2010, which reported on the success of FoldIt players at predicting protein structure on a target that had eluded automatic methods for a few years.
So far as I know, FoldIt is the only computer game so far to have made any real advances in molecular engineering. (The main other contender, EteRNA, though fun to play, does not capture enough in its scoring function for EteRNA play to do anything very useful.)
Disclaimer: my former PhD student, Firas Khatib, was an author on both the FoldIt papers mentioned here, though he was not part of the original enzyme design team.
