See the following image, where I have run auto-structures:
Segment 81 (glycine) is considered to be part of the helix, hence it is counted as an invalid residue.
Curiously enough, if I mutate the segment to asn and then wiggle (at 0.3 ci and then 1.0 ci), then segment 81 is no longer considered part of the helix and can be mutated back to glycine :P.
Shared solution with scientists as "Puzzle 1355 Invalid residue".
I gave the team a heads up to look for it and the link to this - thanks!
I don't know if this is at all related or not, but on 1354 (which lacks Filters beyond Disulphide) I have a Glycine (seg 54) that is part of a sheet and I've manually set it along with its neighbor to loops, have wiggled, shaken, wiggled more etc etc. Then when I click Auto Structures it goes back to being a Sheet. I could understand this if it were making Hydrogen bonds, but it's not, and I can change both 54 and 43 (which is the sheet opposite to it) to loops and not lose the bond. Therefore I can't quite figure out why they need to be sheets, nor why they are auto-structured as sheet.
Again, whether related or not, I'm not sure. :)