Placeholder image of a protein
Icon representing a puzzle

1249: Electron Density Reconstruction 1

Closed since over 9 years ago

Intermediate Overall Prediction Electron Density

Summary

Created
June 20, 2016
Expires
Max points
100
Description

The structure of this protein has already been solved and published, but close inspection suggests that there are some problems with the published solution. We'd like to see if Foldit players can use the same electron density data to reconstruct a better model. Secondary structure predictions (from PSIPRED) are marked on the starting structure, and provide clues about where the protein might form helices and sheets!



Sequence:


SRYTHFLTQHYDAKPQGRDDRYCESIMRRRGLTSPCKDINTFIHGNKRSIKAICENKNGNPHRENLRISKSSFQVTTCKLHGGSPWPPCQYRATAGFRNVVVACENGLLVHLDQSIFRR

Top groups

  1. Avatar for Anthropic Dreams
    1. Anthropic Dreams
    100 pts. 13,648
  2. Avatar for Go Science 2. Go Science 71 pts. 13,647
  3. Avatar for Gargleblasters 3. Gargleblasters 49 pts. 13,625
  4. Avatar for Beta Folders 4. Beta Folders 33 pts. 10,446
  5. Avatar for Contenders 5. Contenders 22 pts. 10,133
  6. Avatar for L'Alliance Francophone 6. L'Alliance Francophone 14 pts. 9,960
  7. Avatar for Void Crushers 7. Void Crushers 8 pts. 9,853
  8. Avatar for xkcd 8. xkcd 5 pts. 9,299
  9. Avatar for Italiani Al Lavoro 9. Italiani Al Lavoro 3 pts. 9,104
  10. Avatar for Deleted group 10. Deleted group pts. 8,572
  1. Avatar for Susume
    1. Susume Lv 1
    100 pts. 13,618
  2. Avatar for frood66 2. frood66 Lv 1 98 pts. 13,541
  3. Avatar for Bautho 3. Bautho Lv 1 95 pts. 13,512
  4. Avatar for bertro 4. bertro Lv 1 92 pts. 10,404
  5. Avatar for LociOiling 5. LociOiling Lv 1 89 pts. 10,304
  6. Avatar for retiredmichael 6. retiredmichael Lv 1 87 pts. 10,124
  7. Avatar for fiendish_ghoul 7. fiendish_ghoul Lv 1 84 pts. 10,107
  8. Avatar for mirp 8. mirp Lv 1 82 pts. 10,103
  9. Avatar for gitwut 9. gitwut Lv 1 79 pts. 10,093
  10. Avatar for grogar7 10. grogar7 Lv 1 77 pts. 10,076

Comments

TomTaylor Lv 1

Clicked to open puzzle 1249 and foldit crashes within seconds. Have tried five times now. Running Windows 10.

Susume Lv 1

Have crashed 4 out of 4 times when opening puzzle. In a new client, once with network timeouts enabled and twice with disabled. Also when switching to 1249 from another puzzle. Puzzle screen comes up, protein is visible, scoreboard loads, then screen goes gray and crash message from Windows.

Win 7 64, 20160609-a37c535a0d-win_x86-devprev

bkoep Staff Lv 1

This seems to be a problem with the "electron_density_panel/threshold" setting in the options.txt file. We're working on a fix for this bug; in the mean time, players can try deleting their options.txt file before starting a new client.

Susume Lv 1

Or if you are comfortable editing your options.txt file, set
"electron_density_panel/threshold" : "0.50"

TomTaylor Lv 1

After comparing the old and newly created options.txt files made the following setting change.

"electron_density_panel/threshold" : "0.500000"

Appears to be working fine now.

gitwut Lv 1

If the structure is known (or almost known) why is the density cloud image so poor? I've played with all the settings and I can't seem to make out a single structure from it.

I'm not a fan of either the Electron Density or Contact Map puzzles, but if the structure is reasonably known, wouldn't a contact map puzzle have been a better choice for this puzzle?

bkoep Staff Lv 1

This density has the potential to look much clearer. When generating this map, we took some steps to limit model bias (a phenomenon in which a poor model can be used to make a deceitfully clear map), but this process also removes some genuine information from the density. Model bias is probably not a huge problem in this particular case, but we wanted to allow Foldit players the opportunity to meet the greater challenge. It is very likely we will repost this puzzle in the future with a refined density map.

In general, an electron density map will contain much more information than a contact map. A model that fits electron density data is much, much more reliable than a model that fits contact data.