For example, will remix, rebuild, wiggle, shake, or idealize
ever add a cis peptide bond to your solution? Can cis bonds
score better than trans bonds in the right environment? Will
any tools (or even wiggle powers like auto, medium, or high)
automatically remove cis bonds from your solution?

I am still curious about the above questions, if
anyone knows the answers to them. A related question 
is, given the right sequence, will AlphaFold ever put 
a cis peptide bond into a Foldit protein?

For more info on cis/trans peptide bonds, see:
https://en.wikipedia.org/wiki/Peptide_bond#Cis/trans_isomers_of_the_peptide_group
https://en.wikipedia.org/wiki/Dihedral_angle#Proteins

Thanks!

Great questions!

In Foldit you can quickly find cis peptide bonds by enabling Show backbone issues in the View Options. Cis peptide bonds are marked with a yellow indicator. You can click the yellow indicator to instantly flip the bond from cis to trans.

Foldit does not have a tool to convert the other direction (from trans to cis) because cis bonds are rare and usually undesirable -- especially in protein design puzzles. I think the easiest workaround to generate a cis bond would be to insert a cutpoint between two residues, and manually rotate one side (180˚ about the axis of the peptide bond) before closing the cutpoint.

Both Remix and Rebuild can produce a cis peptide bond, although this will happen rarely. I don't believe Wiggle will ever convert between cis and trans. AlphaFold predictions will sometimes include cis peptide bonds.

One way to tell cis and trans peptide bonds apart is the 
distances between their alpha-carbons Ca and other backbone 
atoms. Below is a chart of some distances for the peptide 
bond between segments n and n+1:

               approximate
              distances in 
                Angstroms
               cis   trans
Ca[n]-N[n+1]   2.45  2.41
Ca[n]-Ca[n+1]  2.91  3.79
 O[n]-Ca[n+1]  3.63  2.80
 O[n]-N[n+1]   2.25  2.25

One problem with using bands to enforce these in Foldit
is that Foldit doesn't seem to let players or recipes
put bands directly between backbone atoms on adjacent 
segments. If someone can figure out a way to do it, 
please let me know.

It would help if Foldit had LUA commands like below:

str = getcistrans(segnum)
setcistrans(segnum,str)

In these commands, str is a string like 'cis' or
'trans', and segnum is the lower segment number
for the segments making the peptide bond. This
means that getcistrans(5) would be checking if
the peptide bond between segments 5 and 6 was
cis or trans. Meanwhile, setcistrans(7,'cis')
would be setting the conformation of the peptide
bond between segments 7 and 8 to cis.

It would also help if players could select by hand 
two adjacent segments in the protein and then press 
something to switch the peptide bond between these 
two segments from cis to trans or from trans to cis.

Thanks!
Jeff