Re: [AMBER] counter ions in TI calculations

From: Josh Berryman <>
Date: Thu, 13 Feb 2020 08:45:58 +0100

HI Sarah, Sander and pmemd implement the same forcefields, so in simulation
time the convergence should be the same. In wall time pmemd is usually
faster, and pmemd.cuda is much faster.

>>/is there another way to neutralize a protein system?
For most charged sidechains there are alternate parameter sets (variant
residue types that you can name in your initial pdb file) which have
different protonation states. You could pick a few charges on the surface
of the protein and switch them to the neutral versions, and that might even
be physically correct for a solution of some given pH.


On Thu, 13 Feb 2020 at 01:04, Maier, Sarah <> wrote:

> Hi,
> I had a question about using counter ions in pmemd TI calculations of
> proteins. I know that counter ion distributions have notably slow
> convergence in MD simulations for sander, but what about with pmemd? And if
> it is just as slow, how might you remedy this/is there another way to
> neutralize a protein system?
> _______________________________________________
> AMBER mailing list
AMBER mailing list
Received on Thu Feb 13 2020 - 00:00:01 PST
Custom Search