Hello Everyone,
I'm running an MD simulation of a DNA nonamer in a truncated octahedral box with NMR restraints applied only to the terminal base pairs, using AMBER9.
When the two complementary DNA strands are imaged at opposite sides of the octahedral simulation cell at the time when the final restart file is written, the next simulation which uses those restart coordinates as a starting point fails to properly address the periodicity of the restraints on the terminal base pairs of separated strands.
The restarted simulation shows the restraints on the termini pulling those ends together from within the simulation cell, rather than through the periodic boundary. Thus, ripping the complementary strands apart and destroying the structure. It's as though the periodicity of the restraints are "forgotten" when the simulation is restarted. I'm using iwrap=1 because this MD run is one of a series of very long MD runs, and I didn't want to run into a problem with the number of characters needed for coordinate values.
Any suggestions for correcting this problem with restarting these jobs would be greatly appreciated. Thanks!
-- Keith
Here's what the input file looks like:
# Nonamer DNA, 15ns MD run
&cntrl
irest=1 nstlim=7500000 dt=0.002 nscm=1000 t=0.0,
imin=0 iwrap=1 ntx=5 ntxo=1 ntxo=1 ntpr=1000,
ntwr=500000 ntwx=2500 ntwv=0 ntwprt=0 ntr=1 ntrx=1,
ntf=2 ntb=2 dielc=1.0 cut=10.0 scnb=2.0 scee=1.2,
ntc=2 tol=0.00002 ig=83368 ntt=3 gamma_ln=2.0,
temp0=300 tempi=300 ntp=1 pres0=1.0 comp=44.6,
taup=2.0 nmropt=1 scalm=100 ipnlty=1 pencut=1.0,
/
&wt type='TEMP0', istep1=0, istep2=0, value1=300, /
&wt type='END', /
LISTOUT=POUT
DISANG=DNAseq-terminiWConly.rst
END
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Tue Jun 21 2011 - 10:00:06 PDT
