Puzzle 397 is at:
- < 15 version is at http://fold.it/portal/node/989223
The puzzle started on 02/02/2011 and ended on 02/07/2011
"This is the first Quest to the Native Design Puzzle. As mentioned in the blog (http://fold.it/portal/node/989141) we can identify an interaction between a disembodied sidechain and the protein surface and we then want to figure out a backbone conformation that can harbor that sidechain. This quest to the native is based on Puzzle 389: Design the Interface 5 (which you have seen in Gallery Mode). We want to see if this new disembodied sidechain representation (Trp 130) works well for you."
The top score was 9,740 (evolver score by Stardancer).
Several people kindly posted screenshots of their top solutions:
- Stardancer's 1st place evolver solution: tbd
- Steve001's 2nd place evolver solution: tbd
- itskimo's 3rd place evolver solution: tbd
- grabhorn's 4th place evolver solution: tbd
- Bletchley Park's 5th place evolver solution: tbd
- Mark-'s 1st place solo solution:
- Tony Oragami's 2nd place evolver solution: tbd
- grabhorn's 3rd place evolver solution: tbd
- Mr_Jolty's 4th place evolver solution: tbd
- CharlesFortConscience's 5th place evolver solution: tbd
Excerpts from above blog link:
"As some of you may have noticed (http://fold.it/portal/node/989140), we unveiled a new wrinkle in this most recent puzzle: we added implicit rubber bands to position a tryptophan residue in a pocket on the opposite protein interaction partner. You will see these sorts of problems in the months ahead.
The reason we are posting these sorts of puzzles is to help us better design proteins that can inhibit the flu virus. In many cases, we can identify an interaction between a disembodied amino acid and protein surface that is likely to be very beneficial. The trick is to identify a backbone configuration that can harbor that sidechain.
It turns out that we can do this for simple cases where the backbone is a helix or a sheet; but where the backbone is a loop, we are so far unsuccessful at sampling the correct configuration to allow design of new binders. But this is one of the most important classes of problems in protein-protein interactions! For example, many structures of human antibodies bound to proteins of pathogenic organisms fall into this class.
We are relying on you Foldit players to teach us the sorts of strategies we should be attempting for these challenging but very rewarding problems!"
saksoft2 said in response: " options for the next pocket problem I very much enjoy the challenge presented by the FoldIt problems. However, from a pure research standpoint it may be interesting for us folders to have a little extra leeway during the next interface design puzzle.
While you can keep the same format you have now, it may be interesting to see what can happen if we are allowed to grow or shrink the loop to achieve a better pocket fit.
Also, while it is useful in a puzzle like this to have a given point of contact (the tryptophan you banded in for us) an open model may also give interesting results.
So here's my suggestion: for the next interface puzzle, offer two variations of the same puzzle. The first can be just like this one with the fixed length loop and the pre-formed bands. The second would not have the bands and would allow adding/removing elements from the loop we are allowed to manipulate."
beta's reply: " it's tricky... The problem we have noticed about letting players add/remove residues is that everyone always adds the maximum possible number of amino acids. The reason for this is simple: Rosetta gives you a better score if you have more residues.
The new puzzle we just posted, 393: Design the Interface 7: http://fold.it/portal/node/989145
needs work at the terminal end of the protein, so we are able to give you a lot more leeway compared to the previous Design the Interface puzzles.
Despite all the restrictions we had given you, however, we still noticed quite a lot of variation in all the player solutions!"
beta said: " Comments on the new disembodied sidechain representation? Please let us know what you think about the new disembodied sidechain representation. Thanks for your feedback!
Also note that this puzzle will close a day earlier than usual since it shouldn't take much time to work on."
spmm replied: "matching the guide does not seem to score much
Having matched the guide and the sidechains which are unambiguous from the 'shadow' my score moved alarming lower even after rebuilds stabilising etc to improve it.
Having looked at a screen shot of a high scoring solution it is very different to the guide.
So why not just give us the formation shown in the guide and constraints to start with so we can improve from there.
My low score may of course be just my problem but closer to a 'guide' should reult in higher score or one just ignores the guides.
Bit confued by the intent."