3 replies [Last post]
dbuske's picture
User offline. Last seen 2 years 26 weeks ago. Offline
Joined: 09/22/2011
Groups: Beta Folders

No wonder Baker Lab likes the 4 or 5 sheets with 2 helices underneath.
They designed the protein.
I wondered why the folders use that design all the time.

LociOiling's picture
User offline. Last seen 1 hour 49 min ago. Offline
Joined: 12/27/2012
Another "classic of Foldit"
dbuske's picture
User offline. Last seen 2 years 26 weeks ago. Offline
Joined: 09/22/2011
Groups: Beta Folders
Thanks for the reference. My

Thanks for the reference. My fellow team member.

Susume's picture
User offline. Last seen 4 hours 33 min ago. Offline
Joined: 10/02/2011
Scoring selects for the sheets-with-helices designs

I think it is the scoring algorithm (including filters) that leads people to try these designs over and over (and win with them). In foldit, helix segments tend to score highest, followed by bonded sheets, with loops scoring worst (partly because of the number of H bonds per segment in these 3 shapes). For a very long time there was a filter allowing no more than 50% of the segments to be helix. So the strategy that tended to get the most points was 1) use as much helix as the filter allows; 2) keep the loops as few and as short as possible; and 3) satisfy the core filter. If you are limited in your helices and minimizing your loops, the rest will necessarily be sheets. If you only use 2 helices (in order to keep the number of loops minimized), the core filter almost demands that they be on the same side of the sheet wall.

Many players interpret this as two helices followed by 3-5 sheets, or 3-5 sheets followed by 2 helices, but personally I think having at least one sheet-helix-sheet (SHS) area makes the protein more stable and less likely to misfold. Top7 is nice in that its secondary structure sequence is SSHSHSS, so it has two SHS motifs. I generally forgo the long skinny designs (which have the fewest loops) for shorter fatter designs with more sheets and helices, just so I have the freedom to try more different sequences of sheets and helices, and to sometimes put one or more helices on the "back" side of the sheet wall instead of all on the front.


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