[gmx-developers] free energy calculations with restraints

[Shirts, Michael (mrs5pt)]

> The pull code is more general in the sense that it can work not only on
> atoms, but also on COMs.
> But the normal bonded code currently supports a lot more potential types

One possibility is to allow the pull code to access bonded functions.

> and will be more efficient, especially in parallel, since it doesn't
> require extra, global communication for restraining ligands.

This is a good point that I was not aware of.  What is the overhead here?

> We are also planning to clean up and generalize the pull code setup by
> separating the pull groups from the distances and the reaction
> coordinates. That would also make it easier to add angle potentials. But
> I don't know what kind of potentials are required for restraining ligands.

Harmonic restraints in bonds / angles / torsions.  Fairly vanilla stuff.

One thing I didn't quite understand is "The only disadvantage is that the
ligand will not automatically end up in the same periodic image as the
protein."  What are the consequences of this?

Best,
~~~~~~~~~~~~
Michael Shirts
Assistant Professor
Department of Chemical Engineering
University of Virginia
michael.shirts at virginia.edu
(434)-243-1821