AMBER Archive (2008)

Subject: Re: AMBER: Sources of disruption during NVE

From: Gustavo Seabra (gustavo.seabra_at_gmail.com)
Date: Fri Oct 03 2008 - 08:39:05 CDT

On Wed, Oct 1, 2008 at 9:10 PM, Naser Alijabbari <na3m_at_virginia.edu> wrote:
> Hi Gustavo,
>
> I will try to explain what I am doing as best as I can.
> [...]
>>Those simulation times are *very* small, and it's likely that your
>>system isn't even close to equilibrating yet. How large is it? Did you
>>look at any convergence properties to make sure it's OK? It is
>>possible that the differences you see are just the consequence of not
>>having an equilibrated ensemble.
>
> The system has 12154 atoms (10500 water and 1654 protein 2TRX), and the
> pressure, temperature, density and energy all seem to have equiliriated at
> the end of 100ps density adjustment, based on looking at the output file (I
> am using the Perl script provided in the tutorial). Therefore, the
> production run for another 100ps should be equilibriated system.

I still think those simulation times are very small. I wouldn't be
surprised if the differences you mention are just an artifact of the
system not being equilibrated yet. Notice that, depending on the
property you are looking at, it may take longer to reach an
equilibrated ensemble than if you are just looking at total or
potential energies, or any property you can get directly from the
output file, for that matter.

Also, after you bring the system to the desired temperature and
pressure, you still need to give it some time to relax (equilibrate),
before you can do production runs. to judge if you can consider your
system equilibrated enough, you look at the property of interest: it
is roughly the time it will take for the property to stabilize. It's
only after reaching this point that you start the "production run".
Also, remember to throw away the data from the equilibration period,
and consider only the "production" period in the measurements.

>>That's confusing... How different is "different"? In principle, to
>>extract any observable from an md calculation you *must* average it
>>over the simulation time. So, since you mention that the average
>>gives about the same result, maybe they aren't as different... BTW,
>>how did you do this averaging?
>
> I feed my trajectory/.mdcrd file to ptraj to create a mass weighted
> covariance matrix. My understanding was that this matrix is
> created by averaging the variation in position of each atoms for the entire
> time of production simulation.

That seems correct. However, it will not help much if your system
isn't equilibrated yet. (See above.)

>>The discussions you mention are about a completely different thing (I
>>believe). There, what's being discussed is energy conservation, which
>>doesn't seem to be your issue.
> So they are saying tolerance have to be tighter for longer simulation
> periods, and I am guessing tighter tolerance mean longer calculation times.

Actually, the question of energy conservation is only /being noticed/
as you reach longer simulation times. But it is always present, even
if you can't notice in smaller simulations. The tighter tolerance
limits are better is all cases.

> For an NVE simulation w/ no translation removal during the simulation, are
> there any other source of disturbance I am missing? I define translation
> removal, pressure adjustment, and temperature adjustment as disturbances
> because looking at the equations something is being either reset or
> rescaled. So I thought SHAKE falls into that category. I am correct in
> assuming volume is not touched when using NVE?

Translation of the system shouldn't affect any property. Temperature
and pressure regulation do affect the dynamics, and that's the reason
you should use NVE to calculate dynamical properties. As for SHAKE, as
long as the property you want doesn't depend on the vibrations of
bonds to H atoms, you should be OK. Otherwise, you need to turn off
shake for those atoms. If you need flexible water molecules, you can
look into "set default FlexibleWater" in LEaP.

Gustavo.
-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber_at_scripps.edu
To unsubscribe, send "unsubscribe amber" (in the *body* of the email)
      to majordomo_at_scripps.edu