Please see the links below for some questions 
about the Trim and Untrim Tools:

https://fold.it/portal/node/2013141#comment-46429
https://fold.it/portal/node/2013246

Also, could some of the pdb's we are trying to
improve in the Reconstruction Puzzles (see 
https://fold.it/portal/node/2013193 for more
details) need that improvement because the 
software used in their original ED-fitting did 
not include cis peptide bonds? I don't think 
Foldit really lends itself to including cis 
peptide bonds right now, but maybe it should. 

https://fold.it/portal/node/2006311
https://fold.it/portal/node/2006311#comment-46377

https://fold.it/portal/node/995580#comment-24061 
says that cis peptide bonds are quite rare 
compared to trans ones (the trans:cis ratio is
usually about 1000:1, but near prolines it can 
be 3:1). This means that near prolines 75% of
peptide bonds are trans while 25% are cis. 25%
is a pretty big population to ignore. Why not 
make it easier for us to consider cis peptide 
bonds in Foldit puzzles? It could lead to better
results.

https://fold.it/portal/node/995580#comment-24061 says:
---------------------------------------------------
We warn you about cis because, as it turns out, cis 
bonds are very rare in the natural proteins that we 
have observed (1000 to 1 in favor of trans).

There's a small exception to this in the case of 
prolines - when a proline is the top residue in a 
peptide bond, the ratio is about 3 to 1 in favor of 
trans, and consequently we do not create a warning 
about this.
---------------------------------------------------
What does "proline is the top residue in a peptide
bond" mean? Does this mean that of the 2 adjacent 
segments in a peptide bond, the lower #'d one is 
the bottom residue while the higher #'d one is the 
top residue? Are x-pro cis peptide bonds more 
common than pro-x ones (here I list the lower #'d
segments first, and x is any amino acid)? Do
certain amino acids in the x positions make cis
peptide bonds more likely?

I made a list of some recent puzzles, 
how many segments each one had, and 
how many prolines each one had. Then
I divided the # of prolines in each 
by 4 to estimate how many cis peptide 
bonds each should have:

puzzle #segs #prolines #cis  
2167    129    2        0.5   
2158    588   34        8.5   
2155b   588   34        8.5   
2142/3  124    2        0.5   
2137    143    5        1.25  
2128     69    4        1     
2125     95    3        0.75  
2120    128    4        1     
2116     54    5        1.25  
2114    215   11        2.75  
2110     80    6        1.5   
2108    127    4        1     
2100    124    2        0.5   

In this chart, most puzzles should 
have at least one cis peptide bond. 
Also, puzzles like 2167, 2142/3,
and 2100 each have a 50% chance of 
having a cis peptide bond while 
puzzle 2125 has a 75% chance of 
having a cis peptide bond. 

The above makes me think that many 
of our solutions could be improved 
if it was easier for us to include 
cis peptide bonds in them.