[WIP] Add analysis for symmetry corrected RMSD

[hannahbaumann]

No description provided.

[hannahbaumann]

The difference between the old ligand RMSD and the symmetry corrected ligand RMSD on the skipped dataset is quite big and the symmetry corrected ligand RMSD is greater (esp. at frame 5) which seemed weird.
The last test however shows that this is not due to symmetry correction (or the use of the state ligand and not the hybrid ligand), but since we are not doing mass weighting in the symmetry corrected RMSD which has a large effect here due to a Br atom.

Questions:

• Would we want to implement mass weighting for the symmetry corrected RMSD? Do other packages report mass weighted RMSD or not?
• Which "ligand" RMSD do we want to report? So far we had reported the RMSD of the hybrid topology, which, esp. when the number of dummy atoms is large, can lead to a large RMSD, even when the RMSD of the physical atoms might be small.
  • Would we want to report two ligand RMSD, for ligand A and ligand B?
  • Would we want to use physical atoms in the pure end states and hybrid topology for the intermediate lambda windows?

Do you have any thoughts @IAlibay or @jthorton ?

[hannahbaumann]

This test is just here for trouble shooting of the differences between the old RMSD and symmetry corrected. Will remove once things are sorted out.