Re: MMPBSA Energy Calculations using VMD (Post-Processing)

From: Kenno Vanommeslaeghe (kvanomme_at_rx.umaryland.edu)
Date: Mon Apr 14 2014 - 12:24:33 CDT

I'm not sure the PBSA radii for AMBER atom types are accurate when used
with CHARMM charges.

Also, as a reminder, if you use the CHARMM FF for your ligand, you have to
use the CHARMM FF for your protein as well.

Finally, I'd like to point out the "as long as the particle density
reaches approximately bulk outside of the cutoff" assumption is not always
valid. It may not matter for implicit solvent, but I just wanted to have
this on the record so that people who work on other things don't get the
wrong idea. For instance, membranes in explicit solvent are crazy
sensitive to small details in the nonbonded interactions at long range.

On 04/14/2014 08:26 AM, Jason Swails wrote:
> On Mon, 2014-04-14 at 09:56 +0400, James Starlight wrote:
>> Many thanks for suggestions!
>>
>>
>> I'll try to wait AmberTools 14 and test it with the namd's produced
>> trajectory and all input files.
>>
>> One question: can I use parametrization of my system done in
>> charm-like force field instead of amber (including ligand parametrized
>> in cgenff)?
>
> Yes. The "chamber" program has been expanded to support multiple CHARMM
> stream files (assuming you set RTF and parameter sections in a stream
> file for your system via cgenff). The "ante-MMPBSA.py" program now
> supports topology files created with chamber (neither of these are true
> for AmberTools 13). Furthermore, the upcoming chamber release should
> support PSF files created with VMD.
>
>> As I noticed amber need slightly shorter cutoffs for non-bonded terms
>> in comparison to charm indication difference in this terms between
>> both force fields.
>
> Those are just different conventions. The cutoff only applies to the
> van der Waals term (as PME includes the full long-range electrostatics
> over all periodic images). As long as the particle density reaches
> approximately bulk outside of the cutoff, then the long-range dispersion
> correction will correct the energy and virial to eliminate the effect of
> a cutoff and the net force from all particles beyond the cutoff will be
> zero assuming an isotropic distribution of particles at the bulk
> densities. For implicit solvent calculations, you should be using an
> infinite cutoff anyway (as is convention in Amber).
>
> HTH,
> Jason
>

This archive was generated by hypermail 2.1.6 : Wed Dec 31 2014 - 23:22:20 CST