Hi,
> So this set up still no ligand in V1 state. Can you kindly explain why you
> think still maintains ligand internal potentials? I thought this ligand
Because that's the way I wrote the code. Any bonded interaction involving
at least one atom in scmask is made non-lambda dependent, does not
contribute to dvdl and remains unchanged. I think that is what you would
want the behaviour to be in almost any TI calculation.
> Another question is about dvdl_norest, in this kind free energy
> calculation, we usually need restraint the ligand around origin position (
> in protein binding site). So I thought we need set this parameter to no.
> But how to correct this restraint later on?
dvdl_norest does not set or change restraints, it only makes them not
count for dvdl, basically treating them like a bonded interaction of the
ligand. This isn't done automatically because it depends on your
calculation setup if you want this potential counted or not. You still
have to set the restraints yourself (and maybe scale them depending on
lambda)
In the case of:
V0: Protein+Ligand+Ligand_restraint+Water
V1: Protein+Water
we want to set scmask (on V0) to 'Ligand' and dvdl_norest to 'on'. Then
the transformation is equivalent to decoupling the ligand but keeping its
internal potentials and restraint unchanged and not contributing to dvdl.
On additional difficulty is: Since the restraint makes only sense to
define on process 0, it automatically becomes scaled by lambda and we dont
want that. I believe that, as of Amber11, the code automatically corrects
this for you, but check the manual on that.
Kind Regards,
Thomas
Dr. Thomas Steinbrecher
formerly at the
BioMaps Institute
Rutgers University
610 Taylor Rd.
Piscataway, NJ 08854
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Thu Jan 05 2012 - 02:00:03 PST
