Re: NAMD electrostatic interaction Failure.

From: Mehran MB (
Date: Thu Sep 26 2013 - 10:12:16 CDT

Dear Kenno,

thank you for your response and sorry for my ambiguous question. If I want
to say my question again it would be:
by considering 1 for dielectric constant for all atoms in the presence of
molecule of water (explicit model), should one get the same results as
considering 80 for dielectric constant and lack of molecule of water
(implicit Model) !?!

if you think like me that the answer is yes, Now I want to examine that in
NAMD. I was wondering if you could approve my procedure. Here is what I
want to do:

for get about charged residues, Let consider I have box of water and two
elementary charge in that box. Since NAMD use dielectric constant equal to
1 for all atoms since there water molecules (explicit model) I suggest that:

sum over electrostatic interaction between those two elementary charges and
their interaction with water molecules when dielectric constant considered
as 1,
must be equal or at least close to the
electrostatic interaction between just those two elementary charges when
when dielectric constant considered as about 80.

Let me know what you think about it and thank you again for your time and

All the best,


On Wed, Sep 25, 2013 at 8:32 PM, Kenno Vanommeslaeghe <> wrote:

> I'm not sure I understand your post correctly, but you seem to be a bit
> confused about the physics of what you're doing. I'd recommend reading some
> literature and textbooks; see for instance
> showflat&Number=20477<>
> To answer your question, if the interaction energy is that high, the
> residues have to be so close together that there cannot be any water
> molecules between them, so then the dielectric constant of 1 is correct,
> and your (potential) energy seems to be *roughly* in the right ballpark for
> 2 salt bridges. That said, it is indeed disconnected from most real-life
> observables, not because of the dielectric constant, but because said
> observables are typically related to *free* energy *differences* between
> the bound state and a *solvated* reference state (the solvation free energy
> for ions is huge). How to approximate such free energy differences is not a
> trivial question and there exists a large body of literature on this
> subject, and and a large number of methods.
> Oh yeah, for the case of the residues being further apart, there are
> explicit and implicit water models. Setting the dielectric constant to
> anything other than 1 is not a common practice anymore these days (though
> it still has niche applications), so the default of 1 is entirely
> reasonable.
> On 09/25/2013 09:58 AM, Mehran Bagheri wrote:
>> Dear NAMD users,
>> I was trying to calculate electrostatic interaction between two negative
>> and positive charged residues using NAMDEnergy plugin and I got -340
>> kCal/mol when there were close together. 340 kCal/mol is about 574 kT
>> which is huge. Then I found out, dielectric constant is 1.0 which is not
>> correct since I was not consider presence of water molecule. Searching
>> a little bit, I find out NAMD also use 1.0 since it also calculates water
>> molecule interaction. So my fist question is that
>> Is NAMD really use default value 1.0 for dielectric constant?
>> and if yes,
>> Does anyone try to examine this considering 1.0 for dielectric and
>> calculation water molecules will give one reasonable result or not?
>> I did actually, I calculated these two positive residue electrostatic
>> interaction with two negative residue as:
>> E(four residues with itself) - E(two positive with itself) - E(two
>> negative with itself) + E(four residues with water molecules)
>> However it give me even worse result?
>> So Any Idea how I can make this electrostatic interaction more reliable in
>> NAMD would be greatly appreciated?
>> Mehran
>> ps : Switch is 10, Cutoff is 12
>> --
>> Mehran Bagheri
>> PhD. student of physics
>> Department of Physics
>> University of Ottawa
>> 338A MacDonald Hall
>> 150 Louis Pasteur
>> Ottawa, Ontario K1N 6N5
>> Canada
>> Phone: +1 (613) 562-5800 ext. 2817
>> Email: <>_
>> :MB

This archive was generated by hypermail 2.1.6 : Wed Dec 31 2014 - 23:21:43 CST