From: Brian Radak (brian.radak_at_gmail.com)
Date: Mon Apr 15 2019 - 09:51:36 CDT
That's a clever alternative - set the "dielectric" keyword to the AMBER
constant divided by the NAMD constant. There might still be rounding issues
depending on the topology format -- my initial guess is that the above
should work if you are using the AMBER force field with the same AMBER
format input file in both programs.
The choice of PME, lattice, and switching parameters is probably a much
more significant source of numerical error when matching electrostatics.
On Mon, Apr 15, 2019 at 10:05 AM Gumbart, JC <gumbart_at_physics.gatech.edu>
wrote:
> What about changing the dielectric constant, which is a NAMD-accessible
> parameter?
>
> Best,
> JC
>
> On Apr 14, 2019, at 12:47 PM, Brian Radak <brian.radak_at_gmail.com> wrote:
>
> Unfortunately this is not commonly done and requires a bit of an expert
> touch. All internal evaluations in NAMD use the precompiler definition
> COLOUMB inside src/common.h. You will have to change this value and
> recompile.
>
> Note that you might still have a little bit of trouble depending on your
> force field and input file format. The AMBER file format (and the extended
> format for the CHARMM force field) implicitly uses its own value of COULOMB
> (actually, the sqrt of it). A similar thing happens for the CHARMM Drude
> force field in PSF format.
>
> HTH,
> BKR
>
> On Sun, Apr 14, 2019, 2:28 AM Akshay Prabhakant <
> akshayresearch16_at_gmail.com> wrote:
>
>> Hello everyone,
>> I am dong a comparison study for a protein system using AMBER-MD and
>> NAMD, using amber parameters, and for that I need to change the value of
>> NAMD coulomb constant. How can I change the value of the coulomb constant
>> if I have downloaded the NAMD_Linux-x86_64-multicore ?
>>
>
>
This archive was generated by hypermail 2.1.6 : Thu Dec 31 2020 - 23:17:11 CST