OpenMM doesn't compute the Rosetta scoring function. Is that your question, or are you asking something more general about how to compute different sorts of functions?

What do you mean when you say it is "not able to perform the alignment correctly"? What happens when you try? I haven't used MDAnalysis and I don't know how its centering code works. I've always used MDTraj. It has an image_molecules() function that works nicely for centering a protein in the middle...

There's nothing exactly like that. Can you give more details on what you want to implement? What's the functional form? What's the application?

OpenMM 8.1 is now released! You can install it with conda or mamba:
Release notes are at https://github.com/openmm/openmm/releases/tag/8.1.0.

A second release candidate (8.1.0rc2) has been posted. It fixes a couple of bugs related to ATMForce. You can install it with the same command shown above.

The release candidate for 8.1 is now available. You can install it with
Please try it out and let me know whether it works for you. If no serious problems are found, it will become the release in about a week.

In that case, restraining the phi and psi dihedrals might be the best approach.

A common way of doing this is to add harmonic restraints to the positions of some atoms, for example the backbone heavy atoms. There's sample code for doing this in the cookbook.

That expression just linearly reduces the distance. When i is 0 it equals maxDist. When i is 1000 it equals minDist. You can use whatever expression you want for your protocol. To compute the displacement while using periodic boundary conditions, compute it as r=periodicdistance(x, y, z, centerx, ce...

What you have there looks reasonable. You can gradually decrease "dist" as the simulation runs.
Is there something specific that isn't working?