Re: [AMBER-Developers] About switching function in CHARMM and AMBER

From: Ross Walker <ross.rosswalker.co.uk>
Date: Fri, 3 Feb 2017 16:56:09 -0500

My Understand was that Fswitch was the only one that should seriously be considered these days and that is why that is all we have implemented in AMBER. Switching or shifting energy doesn't immediately make sense to me since MD is propogated on the force and adjusting the energy doesn't give you a smooth force term at the cut off (or does it? - need to check the math).

Force shift I've never been comfortable with - it seems to me it would give you behavior that is very dependent on the choice of cutoff - especially for things like VDW where you change things like the mean molecule separation etc by adjusting the cut off and shifting the entire force curve.

All the best
Ross

> On Feb 3, 2017, at 3:51 PM, Thomas Cheatham <tec3.utah.edu> wrote:
>
>
>> I am using CHARMM to do some simulations these days. It seems CHARMM
>> like to use CTONnb and CTOFnb these two keywords to define a switching
>> function. I found AMBER can also use the switching method (but seems
>> only for force, through fswitch keyword). Are the switching method the
>> same in CHARMM and AMBER? Thanks!
>
> Yes, and no. CHARMM implements multiple switching functions including
> force shift, force switch (I think), energy switch, energy shift (I
> think). They like fshift for vdw w/ PME and different "CHARMM" force
> fields choose different options. There is a classic paper by Steinbach
> and Brooks from mid 90's:
>
> New spherical-cutoff methods for long-range forces in macromolecular
> simulation JOURNAL OF COMPUTATIONAL CHEMISTRY
> Volume 15, Issue 7, July 1994, Pages: 667–683, Peter J. Steinbach and
> Bernard R. Brooks
>
> --tom_______________________________________________
> AMBER-Developers mailing list
> AMBER-Developers.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber-developers


_______________________________________________
AMBER-Developers mailing list
AMBER-Developers.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber-developers
Received on Fri Feb 03 2017 - 14:00:02 PST
Custom Search