No replies
jeff101's picture
User offline. Last seen 51 min 54 sec ago. Offline
Joined: 04/20/2012
Groups: Go Science

I wonder if it would be possible to make a new kind of Foldit puzzle like below:

(1) The player is given multiple starting folds for the same protein.
These starting folds are all pretty similar but differ randomly in certain parts.
Perhaps give different starting folds to each player. Perhaps let the same
player try the same puzzle multiple times, each time giving the player a different
set of folds to examine.

(2) No Foldit/Rosetta scores are shown. No wiggles, shakes, rebuilds, etc. are allowed.
Only motions that would not change the Foldit score are allowed.
The player's job is to rank the starting folds by how he/she thinks they should score in Foldit.
When the player is finished ranking the folds, the Foldit scores and correct ranking
for each fold are shown. The player gets points based on how well he/she ranked the folds.

(3) Various viewing methods are available, like show clashes, voids, color by score, hydro, etc.
Rama Maps, Contact Maps, and Electron Densities are shown if applicable. Perhaps let Tab show
scores for each residue. Restrict Recipe use.

(4) This type of puzzle could run after a normal puzzle and use actual Foldit player solutions
from the normal puzzle as its starting folds.

The goal is to get folks used to inspecting structures and recognizing ones that are better
folded than others. If enough folks play this kind of puzzle, you might find certain folds
that most players or the best players rank incorrectly. Perhaps you will find that some
of these incorrectly-ranked folds really should be scored differently by Foldit/Rosetta.

What to do you think?


Developed by: UW Center for Game Science, UW Institute for Protein Design, Northeastern University, Vanderbilt University Meiler Lab, UC Davis
Supported by: DARPA, NSF, NIH, HHMI, Amazon, Microsoft, Adobe, Boehringer Ingelheim, RosettaCommons