I would appreciate any help that you may give... even though it may be
a limitation of the system that i am currently using. I have a very
simple heterodimeric protein that i want to restrict the two c-termini
from moving, or at least maintain a reasonably tight distance from one
another (~43 Angstroms). My system is a gpu cluster that runs
pmemd.cuda.MPI with mpiexec. Amber11. If you need more info let me
know.
I have tried contraining the c-terminal residues in the following way
with the following results:
my first choice:
nmropt=1 / "the gpu version of amber does not support nmropt option"
my second choice:
ibelly=1 / "the gpu versioni of amber does not support ibelly option"
i have been able to restrain the residues but with limited success. I
want them to stay within +/- 1 Angstrom from where they start (there
are reasons for this). i have tried to restrain the terminal 3
residues of each chain but they still drift significantly with time (~
+/- 5 Angstroms). Here is the input file for the first MD run:
Production dynamics with SHAKE,
&cntrl
ntx=1,
ntb=2,
nstlim=5000000,
ntwe=50,
ntpr=500,
nsnb=10,
irest=0,
imin=0,
ntt=1,
taup=0.2,
temp0=300,
dt=0.002,
ntp=1,
ntc=2,
pres0=1,
tempi=300,
ntwx=2500,
ntf=2,
ioutfm=0,
iwrap=0,
ntr=1
/
Restraint of C-terminal residues
250
RES 113 116
RES 236 239
END
END
a restraint of 500 results in an error about something going outside
the box. I have looked at the out file of the run and the correct
residues were selected. I have also looked at the Archives and didn't
find anything that works on my system.
do i have any other choices?
thanks, Worker Bee
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Wed Jun 06 2012 - 18:00:03 PDT
