Secundary structure prediction:
LLLLLLLLLLLLLLLLLLLLLLLLLLLLHHEEEELLLLLLLHHHHHHHLLLEEEELLLEEEEEELLLLEEEEEEEEEEEHHHHHHHHHHHLLLLLLLEELLLLLEEELLLLEEEEEEELLLLLLLLLLLEEEEEELLLLLEEEEEEEEEEEEEEEEELLLLLLLEEEEEELLLLLLEEEELLLLLLLEEEEEEELLLEEEEEEEEEEELLLLLLLLLHHHHHHHHHHHHLLLLLLLLEEEEEEEELLLLLLLLLL
The Zhang Lab link has the "Top 5 final models predicted by I-TASSER". You have to scroll down a bit to see them.
A quick check shows they look a lot like the five structures in this multistart puzzle.
The Select Best Start (SBS) recipe is a classic for this type of puzzle: https://fold.it/portal/recipe/22173
SBS tries all the starting structures. It uses the Lua command puzzle.StartOver: https://foldit.fandom.com/wiki/Foldit_Lua_Function_puzzle.StartOver
Each call to StartOver advances to a new starting structure. The same thing happens when you reset the puzzle with control-r. You'll see a quick pop-up telling you which structure loaded.
If you hover over a segment and hit the tab key, the segment information window opens. This also shows which structure is currently loaded.
All the starting structures have features which don't seem quite right. For example, structure 5 starts with long section of loop which wanders from one side of the protein to the other.
The other structures seem to have similar issues.
"This is a multi-start puzzle, with 5 different starting structures provided by the Zhang structure prediction server"
Indeed the Zhang models we provided to you are the 5 predictions in the link above.
for a protein of this size and complexity. Especially with multiple starts!
https://zhanglab.dcmb.med.umich.edu/I-TASSER/output/S616803/3kg4qm
The above link shows the Predicted Secondary Structure, Predicted Solvent Accessibility, and a ton more information for this protein.
The starting models do not have any assigned secondary structure, because we do not know how accurate the predictions are.