RMSD calculations with correction for sidechain symmetries

[Elaine Meng]

Some RMSD and superposition calculations correct for symmetry (aka "automorphism") by testing chemically equivalent atom pairings for lower RMSD rather than pairing exactly matching atom names only. For example, without this correction the all-atom RMSD between two phenylalanines might be high even though the rings are exactly superimposed because the corresponding atom names are on opposite sides of the rings.

It is harder to do this for arbitrary molecules, since that would require chemical detection, but this suggestion is to cover the standard amino acid residues with standard residue and atom names. I poked around on the web to find out what other programs correct for automorphism in RMS and/or superposition calculations:

---

RMSD calculation symmetry checking
for protein sidechains like Phe, Tyr, Asp, Glu, ... (finite set with std names)
and/or chemically equivalent groups in docked ligands (harder problem)

previous chimera-users post:
​http://plato.cgl.ucsf.edu/pipermail/chimera-users/2009-September/004294.html

software:

ProFit RMS command:
​http://acrmwww.biochem.ucl.ac.uk/software/profit/doc/node15.html

VMD Swap plugin
​http://physiology.med.cornell.edu/faculty/hweinstein/vmdplugins/swap/
I like their default list, which avoids swapping atoms of different elements:
PHE {sym CD1 CD2 CE1 CE2}
TYR {sym CD1 CD2 CE1 CE2}
ASP {sym OD1 OD2}
GLU {sym OE1 OE2}
ARG {sym NH1 NH2}
VAL {sym CG1 CG2}

Rosetta SidechainRmsdFilter.cc
​https://www.rosettacommons.org/manuals/archive/rosetta3.5_user_guide/d1/df6/_sidechain_rmsd_filter_8cc.html

BioSuper superposition web server
​http://ablab.ucsd.edu/BioSuper/
...according to 2013 paper, uses "equivalence of symmetric side chain atoms"
​http://www.biomedcentral.com/1472-6807/13/32

OpenBabel
​https://gist.github.com/baoilleach/974477

Rosetta ligand_dock auto_rms options
​https://www.rosettacommons.org/docs/latest/ligand-dock.html

Autodock rmsd calculation (posting also suggests Pymol does the same)
​http://mgl.scripps.edu/forum/viewtopic.php?f=9&t=611&view=next

OpenEye RMSD tool with -automorph option
​http://sebastianraschka.com/Articles/2014_openeye_alignments_overlays.html#rmsdtool

Schrodinger structutils rmsd module
​http://www.schrodinger.com/docs/suite2012update2/python_api/api/schrodinger.structutils.rmsd-module.html