AMBER Archive (2002)

Subject: Re: who to address this to?...

From: Holger Gohlke (gohlke_at_scripps.edu)
Date: Mon Sep 09 2002 - 11:25:00 CDT

Hi Jessica,

> I have three questions for you and I'm not sure just who they should be
> addressed to. I apologize if I am inconveniencing the wrong people.
>
> 1. It seems that if imin=1 the standard coordinate and energy output
> files are not written. I have imin=1, ntwx=300, and mtwe=50 but the
> corresponding output files are empty.

As far as I know, NTWX and NTWE are only "in action" if you run MD
simulation. In case of a minimization, you may want to set NTPR > 0
(which gives you energy information every NTPR step in "mdout"). For
coordinates, setting NTWR < 0 yields a series of restart files, written
every abs(NTWR) steps. This is probably what you are interested in.

> 2. I want to evaluate the internal energy of a single snapshot in vacumn.
> I would like to use an infinite cutoff for the coulombic interactions. It
> seems that there are two main ways of doing this. One is to implement
> Anal. The other is to either do a single step of minimization of
> simulation in sander. Anal would allow the infinte cutoff but would not
> treat the system according to PME periodic boundary conditions. The
> single step within Sander would not allow an infinite cutoff but would
> treat the system according to PME/pbc. Do you have any further insight
> as to how these two will result in different energies?

If you want to evaluate the energy for a snapshot in vacuum, I do not
see why (and how) you want to use periodic boundary conditions? The
single step in sander can be made to work with "infinite cutoff" by
setting the cutoff to a value larger than the size of your molecule,
e.g. 99999.0 A. In that respect, anal and sander should give very
similar results (apart from the influence of the single step of
minization).

> The current version of MMPBSA uses a single step of minimization. I would
> like to know why this was chosen over a single step of simulation. It
> seems to me that the latter would represent the energy of the ensemble
> more accurately. And beyond the ability to break up the energetic
> contirbutions of individual residues, which I believe sander can do, was
> this switch from anal to sander chosen for any other reasons?

The idea of MM_PBSA is to generate an ensemble first by MD and then
post-process snapshots extracted from that trajectory. Apart from the
(small) influence of one step of minimization, the (internal) energy you
calculate for your snapshot should be the same as the one which you
obtained for that snapshot during MD. In that respect, you work with
your ensemble. One of the reasons to switch from anal to sander is that
sander provides the possibility to use GB models to calculate the polar
part of solvation free energy.

> 3. Finally, I am curious to hear if you have come up with an
> equilibration protocol that seems ideal and applicable to most systems.

I'm not sure if there is an "ideal" equilibration protocol. For
minimization issues, I'd like to refer you to what Dave Case recently
wrote on the amber mailing list. Then, I heat up the system using NVT
simulation while keeping restraints on my solute. After adjusting the
pressure of my system with NPT simulation, I gradually reduce the
restraints and, finally, equilibrate the system without restraints for a
while. This is not very specific information, but I think, the proper
way depends on your system.

Best regards

Holger 

-- 
+++++++++++++++++++++++++++++++++++++++++++++
Dr. Holger Gohlke
Dept. of Molecular Biology, TPC15
The Scripps Research Institute
10550 N. Torrey Pines Rd.
La Jolla CA 92037  USA
phone: +1-858-784-9788
fax:   +1-858-784-8896
email: gohlke_at_scripps.edu
+++++++++++++++++++++++++++++++++++++++++++++