Dear AMBER users,
I am running one step thermodynamic integration calculations of various
molecular systems in explicit solvent (periodic boundary conditions) and
I encountered problem when I use iwrap=1 option with some systems.
The problem is specific to relatively small solute systems where large
percentage of atoms is part of the soft core region. (and the solute
thus significantly differs in size between the two alchemical states)
Since there are really two simulations running for the TI calculations
(each representing the different alchemical state), what happens is that
the solute molecules are at some random point wrapped to the opposite
sides of the periodic box cell. Since there are periodic boundary
conditions, this is essentially fine while the SANDER is running.
However, if I try to restart such simulation from the restart file, it
crashes due to mismatched coordinates.
What seems to me, is that SANDER is not taking into account the periodic
boundary conditions when performing the
"Checking for mismatched coordinates"
at the beginning of the simulation. Coordinates that differ only by
their periodic cell are thus rejected.
Is this a bug or is there some work-around I am missing?
I also point out, that when iwrap=0 is set the calculations run fine
(that is until the coordinates becomes to large for its format which is
really the only reason I'd like to use the iwrap=1 option).
Thank you and best regards.
Miroslav Krepl
Department of Structure and Dynamics of Nucleic Acids
Institute of Biophysics
Academy of Sciences of the Czech Republic
Kralovopolska 135
CZ-61265 Brno
Czech Republic
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Thu Jan 03 2013 - 05:00:02 PST
