Re: Ion type in NAMD Implicit solvent model for ionizing the system

From: Brian Radak (bradak_at_anl.gov)
Date: Wed Sep 13 2017 - 09:59:27 CDT

Quick follow up, now that I've satisfied my own curiosity by looking at
the code. It seems the NAMD implementation inherits some (IMO
ill-advised) idiosyncrasies from the AMBER code on which it is likely based.

1) the ions are /assumed/ to be monovalent - you might be able to change
this, but I doubt it would affect much

2) its not clear that values of the permittivity and Boltzmann constant
are precisely in harmony with the rest of the code (a product of factors
is hard coded), again, not a big deal, but a bit pointless

3) the temperature of the ions is assumed to be either the initial
temperature (i.e., from the "temperature" command) or else 300 K - there
is no coupling with the thermostat. This would only be a big deal if you
are a tad delusional and want to believe the force fields are physically
meaningful at extreme temperature

On 09/13/2017 10:42 AM, Brian Radak wrote:
>
> The generalized Born model in NAMD likely uses the standard
> Debye-Huckel (linearized Poisson-Boltzmann) approximation, which
> augments interactions with the dielectric continuum according to the
> ionic strength. I forget if this includes any effects due to ionic
> size, but it obviously depends on whether the ions are mono- or
> divalent, etc. You'd have to look in the user guide or the code to see
> what actually gets modified.
>
> Note that the approximation is /very/ poor for large concentrations,
> greater than a few hundred millimolar, say.
>
> HTH,
>
> Brian
>
>
> On 09/13/2017 10:26 AM, Chitrak Gupta wrote:
>> From what I understand, salt concentration becomes irrelevant for
>> implicit solvent simulations, because implicit solvents assume
>> infinite dilution. Note that you don't have a periodic box. Even if
>> you did have ions I am guessing they would fly away as soon as you
>> start the simulation. In my understanding, this is one drawback of
>> implicit solvent methods in that you cannot choose an ionic strength.
>>
>>
>> Hope this helps,
>> Chitrak.
>>
>> On Wed, Sep 13, 2017 at 1:49 AM, Roshan Shrestha <roshanpra_at_gmail.com
>> <mailto:roshanpra_at_gmail.com>> wrote:
>>
>> If you use VMD, you can add KCl by using extensions/
>> modelling/add ions.
>>
>> On 9/13/17, Susmita Ghosh <g.susmita6_at_gmail.com
>> <mailto:g.susmita6_at_gmail.com>> wrote:
>> > Dear NAMD users,
>> >
>> > In NAMD implicit solvent model how to specify the type of ION
>> for ionizing
>> > the system? If I use only the "ionConcentration" command then
>> by default
>> > it will take NA+ and CL- ions. Now I want to use KCL instead of
>> NACL to
>> > ionize the system. Is there any any way to do that with
>> Implicit solvent
>> > model?
>> >
>> > Thanks in advance!
>> >
>>
>>
>> --
>> Roshan Shrestha
>> M.Sc (Physics)
>> Central Department of Physics, Tribhuvan University
>> Kathmandu, Nepal
>>
>>
>
> --
> Brian Radak
> Postdoctoral Appointee
> Leadership Computing Facility
> Argonne National Laboratory
>
> 9700 South Cass Avenue, Bldg. 240
> Argonne, IL 60439-4854
> (630) 252-8643
> brian.radak_at_anl.gov

-- 
Brian Radak
Postdoctoral Appointee
Leadership Computing Facility
Argonne National Laboratory
9700 South Cass Avenue, Bldg. 240
Argonne, IL 60439-4854
(630) 252-8643
brian.radak_at_anl.gov

This archive was generated by hypermail 2.1.6 : Sun Dec 31 2017 - 23:21:40 CST