AMBER Archive (2008)

Subject: Re: AMBER: PMEMD and polarized force field ff 02

From: David A. Case (
Date: Thu Jul 03 2008 - 10:10:06 CDT

On Thu, Jul 03, 2008, Abhishek Singh wrote:

> I have a situation that I am simulating rna molecule in a TIP3PBOX using
> ff02 force field and salt libraries ionsjc_tip3p.

Wow. Overall I think you need to proceed more slowly and be sure you
understand basic simulations before trying such exotic things.

First, as Tom has already pointed out, the ionsjc stuff is designed for
non-polarizable potentials; no testing has been made with ff02. (Others can
chime in, but not much testing of ff02 for RNA has been done at all, as far as
I know.)

> While equilibration there
> is a protocol PMEMD that has input file as shown below. Now if i use the
> flag ipol=1

PMEMD does not support polarizable potentials. Be sure to read the Users'
Manual for a list of things like this.

> Kindly suggest how can i run md without PME.

Why do you want to do this?

> PMEMD input file is
> heat/dynamics holding RNA fixed
> &cntrl
> imin=0,
> ntxo=1, ntrx=1,
> cut=9.0, tempi=100.0, ntwprt=0,
> ntpr=100, ntwx=500, ntwe=500,
> nstlim=10000, temp0=300.0,
> nscm=0,nsnb=10,
> ntc=2, ntf=2, tol=0.00001,
> tautp=0.2, t=0.0, vlimit=20.0, comp=44.6,
> ntx=1, irest=0, dielc=1.0, ibelly=0,
> ntb=2, ntp=1, pres0=1.0, taup=0.2,

Don't do this! You have to equilibrate first with constant volume (to the
desired temperature, *then* turn on constant pressure. You should set tautp
and taup to values at least five times larger than you have here.

> scee=1.2, dt=0.002,

Using such a big time step is dangerous, especially for an initial
equilibration step.

> ntt=1, ntr=1,ipol=1,
> &end
> 200.0

This is too large, and can actually cause problems with dynamics. Try a value
about 50 times smaller.

Be sure you have completed and analyzed some simulations of your RNA sequence
with standard potentials and protocols, before trying out things like
polarizable potentials.

...good luck....dac

The AMBER Mail Reflector
To post, send mail to
To unsubscribe, send "unsubscribe amber" (in the *body* of the email)