Re: AMBER: single or dual topology for TI?

From: David Mobley <dmobley.gmail.com>
Date: Sat, 21 Oct 2006 13:58:41 -0700

Holly,

On 10/21/06, Holly Freedman <freedman.phys.ualberta.ca> wrote:
> Thanks , David. Should I follow a similar procedure when creating atoms, ie when going from OCH3
> to OCH2CHCH2, should I first create atoms without charges from dummy atoms, and then in another
> cycle add the charges to new atoms?

Right. Sorry, I didn't explain the reason for this. If you have
nonzero charges on atoms which are being inserted/deleted, unpleasant
things can sometimes happen. In particular, atoms which are being
inserted/deleted can occasionally come extremely close to other atoms
(since, of course, you're turning on/off their Lennard-Jones
interactions). If those other atoms have the opposite charge, you run
the risk of having a "fusion" sort of event, where two point charges
of opposite sign overlap. This tends to have unfortunate consequences
for MD simulations, for obvious reasons. Hence, it probably doesn't
particularly matter whether you're inserting or deleting things -- you
just want to make sure the things you insert or delete are uncharged.

David
-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber.scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo.scripps.edu
Received on Sun Oct 22 2006 - 06:07:38 PDT
Custom Search