Hi, all.
Is it possible to create two types of extra point charges in AMBER?
I want to improve molecular electrostatics for a single molecule by
adding atomic dipoles and quadrupoles. It is important for me to use
GAFF in the same molecule where I put atomic multipoles, and, I guess,
there is no way to use atomic dipoles and quadrupoles (as in AMOEBA)
with GAFF (Correct me, please, if I'm wrong). To avoid this, I'm
trying to simulate dipole and quadrupole with 2 extra point charges (+
atomic charge). The problem is that these two extra points are
different: one is "inside" R-X bond and the other is outside. I need
different atom types to set different valent angles (0.0 for "inner"
EP and 180 for "outer").
Amber manual states that extra points is "atoms with "EP" in their
names". So, I see no problem in using EP as names for all extra points
and creating 2 atom types (EP and LP or EP and ZP to avoid conflicts
with ff99SB's LP) for different positions.
When I have only one EP in "outer" position, it is treated as EP and
everything works fine.
Next, I want 2 different EPs.
I use frcmod ANGL section like this to have different EP positions:
ANGL
XX-br-LP 150.0 180.00
XX-br-EP 150.0 0.00
Both EP and LP atoms have "EP" as their atom names, but tleap seems to
recognize extra point only in EP and asks for angle parameters for
ca-br-LP and EP-br-LP. Ok, I add these parameters and get my prmtop
and prmcrd. Minimization in sander.MPI runs fine, but during it these
extra points are moved from their initial positions, so one is located
as "outer" extra point and the other's bond is perpendicular to R-X
bond (but bond angle should be 0.0 acccording to frcmod). Moreover,
during MD run both extra points are not kept in their ideal geometry .
"Outer" extra point seems to be fixed in space and bromine atom
oscillates somewhere nearby. "Inner" extra point spins around bromine
atom. May be interesting to note this fact: when I set ideal extra
point bond length to 0.1 Ang (to have "point" multipoles) sander
crashes with segfault somewhere in the end of the first MD iteration,
and when bond length is set to 0.5 Ang it seems to work fine, except
extra points are not treated as extra points.
Can anybody tell me what I'm doing wrong? Is it possible to have two
types of extra points in sander?
With Best Regards,
Oleg.
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Wed May 16 2012 - 03:30:03 PDT
