AMBER Archive (2009)

Subject: Re: [AMBER] wrong estimation of hydrogen velocities when using SHAKE?

From: Jose Borreguero (borreguero_at_gmail.com)
Date: Wed Mar 11 2009 - 12:28:21 CDT


Dear David,

I ran simulation without SHAKE and hydrogens did show correct kinetic
energy.

I assume that since SHAKE reduces the number of degrees of freedom by one
due to the constrain, then maybe for hydrogens Ek=kT instead of (3/2)kT. I
computed Ek for each hydrogen in the simulation with SHAKE, and using Ek=kT
I got T about 6% higher than system temperature. Pretty close.

Thanks for your suggestions.

On Wed, Mar 11, 2009 at 7:08 AM, David A. Case <case_at_biomaps.rutgers.edu>wrote:

> On Tue, Mar 10, 2009, Jose Borreguero wrote:
> >
> > I am running NPT simulations and apparently the speeds of the hydrogens
> are
> > systematically underestimated roughly by a factor of sqrt(2). I have
> checked
> > with two different proteins and two different temperatures, having same
> > results
> >
> > I am using the SHAKE algorithm, so I ignore how hydrogen velocities are
> > estimated in this case.
>
> Can you give more details:
>
> 1. How did you find out the velocities of the hydrogens?
> 2. How did you decide what it "should" be?
> 3. What happens if you turn SHAKE off? (As you point out, it is
> complex to analyze velocities when constaints are present.)
>
> ...dac
>
>
> _______________________________________________
> 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