AMBER Archive (2009)

Subject: Re: [AMBER] Changing Force Fields in "Midstream"

From: David A. Case (
Date: Wed Mar 11 2009 - 17:36:56 CDT

On Wed, Mar 11, 2009, Wei Zhang wrote:
> Try use ions94.lib to replace ions08.lib.
> The problem is in ions08.lib, the atom type for Magnesium was set
> to Mg+ which is wrong, it should be MG. ions94.lib does not have this
> problem.

Well, please be careful here. I don't know what ions you have in your
systems (both monovalent and divalent)? If you monovalent ions, and if
you want the newer parameters for those (see comments below), you can't
just go back to ions94.

Arguably, one problem is that ions08.cmd (and hence, ions08.lib) define
an "Mg+" ion, which is both poorly named (since it is really +2), and
which has no corresponding force field terms in the ions08 frcmod files.
(I'm proposing to eliminate the "Mg+" atom for the next release of
AmberTool, and re-add it back as "Mg2+" if or when we have reasonable
parameters for it -- comments welcome.)

If the original poster wants to mix new monovalent ion parameters with
old MG parameters, it will probably be necessary to do a bit of
hand-editing of either cmd or lib files, since this is not a pre-built

Which then leads to the question for tec: should people be using the new
ions08 stuff for DNA/RNA with ff99bsc0? If so, why does leaprc.ff99bsc0
still load ions94.lib? If not, why not?


AMBER mailing list