to optimize it or make own folds?
Make totally fifferent aligement or "just" optimize existing one?

I looked up Electron Density Trace and I got the impression that the method predicts where each AA probably is with some margin for error. So, does that mean we should try not to change their locations too much? Is there any chance you can show us the original results with the probabilities?

Indeed, the data indicates that this model is close to the native geometrically but you can see that it needs a lot of work and optimization.

So drastic restructuring moves should not be necessary, but finding the native low-energy well should still be a challenge!

I will see if we can provide you with some of the experimental data, if not on this puzzle we can try it with future puzzles.