beta_helix Staff Lv 1
For the puzzles in the CASP9 Refinement category, the CASP organizers have seen the native solved structure already and picked a submitted prediction that is closest to the native. They give us this model and tell us which regions are incorrect, making it easier to focus on which areas of the model to refine.
For 332: CASP9 Refinement Puzzle 2, they told us that "in particular two loops around residues 41 and 55 are substantially off".
In the image below, the starting model given to us is shown in black with the recently released native in blue. The loops they told us about are indicated in red (these were shown in note mode in the puzzle). The 5 other models shown in the figure were the top scoring models from the top 5 Foldit groups for puzzle 332. These models barely changed the indicated region (focusing more on the loops in the bottom left).
We did not submit these top scoring models and instead had to submit much lower scoring Foldit predictions that had refined the indicated loops.
To insure that the regions we have been told are incorrect get sufficiently refined, we are adding conditions to these areas of the protein. You score will not count until you have changed those regions (indicated in note mode) by a certain amount (shown in the conditions tab under your score).
Foldit performed very well in the refinement category at CASP8 (the TR targets in: http://fold.it/portal/node/729520) so we want to insure that you all do as well at CASP9!