I think the problem is what I suspected and also what dave case
sugested- you are changing the refc reference coordinates each
time, so every time you restart there is a jump in the restraint
energy. this also allows a drift in the structure, it is not the same
as what you would get with 1 long run using a single reference coordinate set.
you should probably keep the same reference coordinates for each run.
carlos
----- Original Message -----
From: yen li
To: amber.scripps.edu
Sent: Wednesday, July 14, 2004 12:44 PM
Subject: Re: AMBER: Implicit MD problem
Hi Carlos,
For 1 ns of md, the number of steps involved are 500000 with the time step dt=0.002.
The input file is
&cntrl
imin=0,
ntx=5, irest=1,
ntwx=5000, ntpr=200,
ntf=2, ntb=0,dielc=4,cut=300.0,
ntr=1, nscm=200,
nstlim=500000,dt=0.002,temp0=300,tempi=300,
ntc=2,ntt=1,tautp=2.0
&end
&ewald
eedmeth=5,
&end
restraint 1
25.0
RES 9 17
RES 22 27
RES 32 43
END
END
I am also writing the script in which i am repeating the 1ns of the simulation using the above file, so many times as the number of nanoseconds for which i want the simulation.
1st ns of production phase
sander -O -i mdprodIm.in -o prot1.out -p prot.top -c eqa.restrt -r prot1.restrt -x prot1mdcrd -ref eqa.restrt
2nd ns of production phase
sander -O -i mdprodIm.in -o prot2.out -p prot.top -c prot1.restrt -r prot2.restrt -x prot2mdcrd -ref prot1.restrt
similarly for more runs.......
where
prot.top is the topology file which will be same throughout.
mdprodIm.in is the input file for the production phase given above which will again be same for each of 1 ns simulation.
eqa.restrt is the restrt file from the equilibration phase, here it is being used as coordinate file
prot1.restrt is the new restrt file for 1st ns which will be used as the coordinate file for the 2nd ns of simulation.
prot1mdcrd is the saved coordinate file.
I hope now you can follow how exactly i am doing. I'll appreciate your critical remarks.
Thanks
Carlos Simmerling <carlos.ilion.bio.sunysb.edu> wrote:
It is very hard to help without seeing the actual files that you used.
If your batches were 1ns, why does this input file say nstlim=500000?
How did you get it to stop at exactly 1ns?
Also, please send the scripts so we can see how you did the restarting
and what was used for the refc for restraints, etc.
Please include them in the mail as plain text, since I do not open Word
documents in email.
----- Original Message -----
From: yen li
To: amber.scripps.edu
Sent: Wednesday, July 14, 2004 7:01 AM
Subject: AMBER: Implicit MD problem
Hello amber users,
I am doing the md implicit solvent md simulations on a small protein by putting restraints on the helices, using amber7 suite of programs. The input file for the production phase is given below
&cntrl
imin=0,
ntx=5, irest=1,
ntwx=5000, ntpr=200,
ntf=2, ntb=0,dielc=4,cut=300.0,
ntr=1, nscm=200,
nstlim=500000,dt=0.002,temp0=300,tempi=300,
ntc=2,ntt=1,tautp=2.0
&end
&ewald
eedmeth=5,
&end
restraint 1
25.0
RES 9 17
RES 22 27
RES 32 43
END
END
I am doing the simulation in baches of 1 ns each.The problem is that the temp v/s runlength, total energy v/s runlength and pot energy v/s run length plots shows kinks or drops at the points where i restart the next batch of the production phase. The plots for the 16 ns are attached.
Is there any problem in the restarting of the simulation.
--------------------------------------------------------------------------
Do you Yahoo!?
Take Yahoo! Mail with you! Get it on your mobile phone.
------------------------------------------------------------------------------
Do you Yahoo!?
New and Improved Yahoo! Mail - 100MB free storage!
-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber.scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo.scripps.edu
Received on Wed Jul 14 2004 - 18:53:01 PDT
