> > > Part of a residue I am parameterizing is a square planar metal
> > coordination
> > > complex. I have some questions about the improper torsions that LEaP
> > chooses
> > > to apply. For the purpose of discussion, I have simplified the structure
> > so
> > > that the central atom (a metal atom) is labeled C, and the coordinated
> > > ligands are the points of a compass (clockwise N, E, S, W):
> >
> > > N
> > > |
> > > W - C - E
> > > |
> > > S
> >
> > > These are my improper parameters:
> >
> > > IMPR
> > > E-S-C-W 0.2 180.0 2.0 These 2 impropers are
> > > applied by LEaP
> > > N-S-C-W 0.2 180.0 2.0 .
> > > S-E-C-W 0.2 180.0 2.0 These 4 are not
> > applied
> > > by LEaP
> > > S-N-C-W 0.2 180.0 2.0 .
> > > S-W-C-E 0.2 180.0 2.0 .
> > > S-W-C-N 0.2 180.0 2.0 .
> >
> > > I verified that the first two impropers are applied by examing .C in
> > rdparm:
> > > B 2901: 0.200 3.14 2.0 :25.E :25.S :25.C :6.W
> > > (768,760,758,168)
> > > B 2902: 0.200 3.14 2.0 :25.N :25.S :25.C :6.W
> > > (776,760,758,168)
> >
> > > C is square planar; W and S are trigonal planar; N and E are tetrahedral.
> >
> > > ***From reading the list, I have gathered that assigning impropers is
> > > complicated. Many references are made to LEaP recognizing "planar"
> > > topologies and applying an improper (if parameters are given, otherwise a
> > > warning is generated). How is the "planar" topology recognized? Simply by
> > > finding tricoordinate atoms? Can you explain how LEaP decided to apply
> > some
> > > of these impropers to my tetracoordinate atom?
> >
> > Planar topology as such is not recognized by leap; as you
> > have guessed it examines a list of tricoordinate centers
> > to see if any of the permutations match an improper parameter.
> >
> Can you explain why LEaP assigned some impropers to C, a tetracoordinate
> atom? Two impropers were applied with C as the 3rd/central atom. My
> intuition is that it has something to do with the trigonal planar nature of
> the S and W atoms.
The central atom of any tricoordinate system is always the
third one in the improper. You would have to read leap code
to see how the tetracoordinate center is handled as multiple
tricoordinate terms. It may even be worth a detailed look,
adding print statements to the code, since this is not a common
case.
Ideally I think leap might avoid cases like NSCE where 'bonds'
superimpose, but I didn't think of this case when working on the
code for impropers eons ago.
Bill
> > ***Additionally, looking at the N-S-C-W improper (2nd in the list above):
> > > N-S-C looks like an unstable choice to define a plane, since the three
> > > points are (should be) collinear. Even a slight deviation of C from the
> > > plane would result in a large angle between the planes defined by the
> > > improper. Could this be problematic?
> >
> > You might want to avoid such combos - it could call for unfortunate
> > ripple effects as you create new atom types and add the other params
> > for them though, so I'd try without doing this at first.
> >
> > I'll keep this in mind, but I think I've avoided this, for now. The W atom
> is the only one of its type in this system.
> > > ***Also, I realize I can add the other impropers to the prmtop by hand,
> > but
> > > can this lead to strange behavior from sander?
> >
> > Depends on how well you do it :-)
> >
> I'll try hard then. Thanks!
> Robert
> > Bill
> >
> > > The square planar structure of the coordination complex is represented
> > > sufficiently well with or without these impropers, but I would like to
> > > understand what's going on behind the scenes. Apologies if this has been
> > > answered before, but I did a fairly extensive search of the mailing list
> > > without coming across a satisfactory answer.
> >
> > > Cheers,
> > > Robert Elder
> > > PhD Candidate
> > > University of Colorado - Boulder
> > > _______________________________________________
> > > AMBER mailing list
> > > AMBER.ambermd.org
> > > http://lists.ambermd.org/mailman/listinfo/amber
> >
> >
> > _______________________________________________
> > AMBER mailing list
> > AMBER.ambermd.org
> > http://lists.ambermd.org/mailman/listinfo/amber
> >
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Tue May 25 2010 - 11:00:08 PDT
