AMBER Archive (2009)

Subject: Re: [AMBER] Density in MD simulation

From: Ibrahim Moustafa (I.moustafa_at_psu.edu)
Date: Fri Apr 17 2009 - 18:26:46 CDT

Hi Dr, Carlos,

  Thanks for your reply. The density is expected to be slightly higher than
water for protein solution. I got concerned because the system is very close
to previous simulations where the density was fluctuating around values
nearer to 1.000 and I'm using the same protocol for the simulation.
 It should be mentioned that, in tutorials as well as in many published
simulations, authors, usually, report that the density fluctuates around
values not that far from 1.00 (definitely < 1.04). Also, personally, I heard
from some modelers that for a good simulation run the density should be very
close to unity!. I used to argue against that; I think what is more
important is the stability of density during simulations rather than its
exact value.

  Thanks,
 Ibrahim

   

On 4/17/09 12:10 PM, "Carlos Simmerling" <carlos.simmerling_at_gmail.com>
wrote:

> if you are equilibrating the pressure, not the density, then it may be
> completely fine that the density is slightly high. it depends on
> solute, etc. can't say much more without knowing your solute and
> whether it should be slightly more dense than water.
>
>
>
> On Fri, Apr 17, 2009 at 12:09 PM, Ibrahim Moustafa <I.moustafa_at_psu.edu> wrote:
>> Dear Amber users,
>>
>>  Iım trying to do MD for a protein solvated (TIP3PBOX) in a truncated
>> Octahedron cell (12 A distance from protein atoms to the edge) using Amber10
>> and ff99SB.
>> Currently, Iım at the equilibration stage. I noticed that the density is a
>> bit higher than it should be (1.0430, a little bit off from 1.0000). The
>> following  input parameters are used:
>>
>> -----------------input file-----------------------
>>  &cntrl
>>   imin=0, irest=1, ntx=5,
>>   ntpr=1000, ntwx=1000, ntwr=1000,
>>   cut = 9.0, nsnb = 10, scnb = 2.0, scee = 1.20,
>>   nstlim=350000, dt=0.001,
>>   ntt = 1, tautp=2.0, taup=0.2, temp0=300.00,tempi=300.00,
>>   ntb = 2, ntf = 2, ntc = 2, ntp = 1,
>>   nscm = 500,
>>  &end
>>  &ewald
>>    use_pme=1,
>>  &end
>> /
>> -----------------------------------------------------
>>
>> I tried to use different distances (shorter and longer) when solvating the
>> system and the density behaves the same way, always reach the point where
>> fluctuating around 1.0400. I wonder if there is a way to bring the density
>> closer to 1.0000. In previous simulations for slightly different systems,
>> usually, I have the density fluctuating around 0.998 & 1.020.
>> Iım a little worried about the artifacts in the simulation because of the
>> density value. Actually, Iım not sure how bad the effect could be!
>>  Any comment, suggestions would be appreciated.
>>
>>  Googling and looking through AMBER archive could not help to find an answer
>> to my question.
>>
>>  thanks,
>>  Ibrahim
>>
>>
>> _______________________________________________
>> AMBER mailing list
>> AMBER_at_ambermd.org
>> http://lists.ambermd.org/mailman/listinfo/amber
>>
>
> _______________________________________________
> AMBER mailing list
> AMBER_at_ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
>

_______________________________________________
AMBER mailing list
AMBER_at_ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber