Game change suggestions, thoughts, questions

Case number:699969-995544
Opened by:Steven Pletsch
Opened on:Monday, July 15, 2013 - 07:43
Last modified:Wednesday, July 17, 2013 - 09:17

Finally get back into the swing of things and been thinking a bit about how the game has evolved the past couple of years. I wanted to offer up a couple of thoughts, mostly in the form of questions because I don't know if they would be effective solutions or not:

I think freestyle design types of puzzles are a lot of fun, but I wonder if they are producing the desired results ? Since the highest scoring solutions seems to be helix only designs, I wonder if reducing scores for helix bonds slightly, and increasing slightly the sheet, loop, and sidechain bonds for these types may yield more diverse results ?

I also wonder if it would be possible to create semi-targeted freestyle design puzzles, like simply providing a locked structure, and a separate freestyle design with a large degree of variance in length, say starting with 50 segments, and allowing a variance of +/- 40, and seeing what people come up with for interactions. Or is this ineffective or just too general ?

and finally, on the beginner puzzles, rather than having a repost of something, I wonder if it would be more effective to offer a simple, but unsolved protein structure. I think the real hook to this game is being able to have a shot at "making a difference" and that a lot of new people may be disillusioned by spending hours on a puzzle only to find out it was already done before, or doesn't really "help" in the way they expected.

Just some thoughts I wanted to share ;)

Thanks for all you do, I <3 FoldIt

(Mon, 07/15/2013 - 07:43  |  3 comments)

spmm's picture
User offline. Last seen 11 weeks 6 days ago. Offline
Joined: 08/05/2010
Groups: Void Crushers

May I unpack this a bit for simplicity:
Freestyle - types
__the denovo with a sketch of secondary structure and fixed residues - known chain but no known pose
---the specific length chain, no SS and all residues can be mutated and any pose is possible, these seem to be mostly symmetry puzzles
--- design puzzles, these vary from limited, to providing the ability to add residues and mutate a proportion of the residues

The denovo and design puzzles seem to me at least to have a specific goal which is well explained, any new player would hopefully feel that these are 'important'.

The fully mutable chain I am not really clear on the purpose of these, we did have a few non mutable sym puzzles a while ago which were fun, but open to gaming; they do allow full expression and who knows what people will come up with.
Strangely I find these really interesting to do but I never have an expectation of a high score as the big helix or engineered precision gets a bit tired after a few times, I simply assume that anything I do in that space (residues, pose) will have perhaps existed for a nanosecond in a 'real' protein or may be quite new but unlikely to survive in the wild of the lab.

The challenge is still there, getting the score of the construct above the 'game' scores requires a lot of effort, and a perhaps unfounded belief in the rightness of the solution at that time. I am not sure how many existing proteins there are in the world and how well they would score in foldit. Perhaps a 'the residue chain you have selected is a [insert pdb name here] may be a useful filter.

It did also cross my mind that in the total freestyle one may construct a protein of such a score that would make it zoom into the wet lab and develop into the next big thing :D

People of course differ in their learning styles, I would think that most would appreciate the beginner puzzles as a way to get their chops down, after the brief tutorials, there is no restraint on doing any puzzle and anyone can leap into any science puzzle if they don't feel that the beginner puzzles are for them.

Joined: 05/26/2008
Groups: Hold My Beer

For the design puzzles I am thinking more of having a protein set up to interact with, we have had similar puzzles in the past where 2 are presented, 1 is locked and the other is partially locked with a small area workable to find an interface. What i am thinking on threse is more having a locked protein, pointing out the area to interface with, and then allowing a full from scratch design with no locked areas for the second protein, so we can have a crack at completely designing something that interfaces with it.

As far as the new player puzzles, I think they are good the way they are, I am just wondering if they may stir up more interest if they were targeted towards a specific goal. I know that foldit has had many players come in and leave again after a short period. I am thinking if newer players had something they viewed as very important, it may help with new player retention

spmm's picture
User offline. Last seen 11 weeks 6 days ago. Offline
Joined: 08/05/2010
Groups: Void Crushers

Ah that sounds good, the highly locked ones can be tedious to play, I think that they are locked so that we are really not able to design fully from scratch as the science has done that already and they are looking for ideas in certain spots in particular. That was my understanding from the chats but I may be quite wrong :)


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