AMBER Archive (2009)

Subject: Re: [AMBER] Boundary Condition Problem

From: case (case_at_biomaps.rutgers.edu)
Date: Sat Jul 04 2009 - 08:45:45 CDT

On Sat, Jul 04, 2009, Pepe` Falahat wrote:
>
> I am trying to use Amber 10 to perform energy minimization and MD on a
> system consisting of Guanine tetrads. I have consulted the manual
> extensively and followed the online tutorials, but I've still been having
> difficulty making the simulations work with periodic boundary conditions. My
> main questions are as follows:
>
> 1) How would I create a periodic boundary cell with a non-rectangular
> lattice?

You would load a pdbfile into LEaP, use set Box to set up a rectangular
lattice, then hand-edit the coordinate file, changing the last line to contain
the values of a,b,c, alpha,beta,gamma that you want. This path assumes that
the initial pdb file has all of the waters or other solvent molecules you
need.

Note that periodic simulations in Amber require a largish unit cell, at least
18-20 Ang. in each direction.

>
> 2) Is it even possible to minimise a system with periodic boundaries (ntb =
> 1), without the need of an explicit solvent? By 'minimise a system' here I
> mean minimise the atoms AND the lattice vectors of the periodic boundary.

It's hard to see how a periodic simulation without solvent makes sense, and
Amber does not support periodic GB simulations. Also, there is no facility to
minimize lattice vectors -- do you really want a 0K result? More usual is to
use contant pressure MD to optimize (at some finite temperature) the *lengths*
of the lattice vectors. Amber as no facility to change the lattice angles.

>
> 3) And finally, would I be able to do NVT molecular dynamics with a system
> such as I've just described?

Yes.

....dac

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