Hello,
I am wondering how forces with jar=1 are actually calculated in case of
a distance restraint between two atoms. The way I understand it, a
'correctional' force is dynamically calculated with each MD time step
and applied in order to force the distance d between two atoms on a
certain trajectory d(t), which is a linear relation in this case.
In sander's nmr.F90, I've found `fcurr = -2.0*rk2*(rint-r2)` (line 3778
for Amber 12).
First of all, I am wondering about the direction of the force. I'm not a
Fortran expert, but assuming all of the variables in the code line above
are scalars, only the force magnitude is calculated here. The direction
should be given by the gradient of the according potential. Where/how is
the direction of the force calculated?
Futhermore, I am wondering on which of the two atoms the force is
exerted. Is it calculated and applied independently for both atoms with
half the force constant? Or is it applied only to one atom? If the
latter is true, is it applied to the first or second atom defined via
e.g. `iat=485,134` (in other words: which atom is the fixed end?)?
Then, in the classical sense of F = -kx, rk2 is only k/2 (due to the
factor 2 in the code line above, as already clarified here:
http://archive.ambermd.org/200904/0162.html). In the Umbrella Sampling
section in the AMBER manual there is this remark: "Note that the
conventionally defined force constant is twice the value rk2". My point
is that this should also be clarified in the SMD section, otherwise
people may publish force constants different from those that they
actually used.
Thanks for your help with understanding SMD in Amber!
Jan-Philip
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Mon Dec 10 2012 - 03:00:08 PST
