From: Peter Freddolino (
Date: Mon Oct 26 2020 - 19:16:57 CDT

Hi Kacey,
I'm afraid that I don't really understand the question. Your question
states that "The values calculated are exactly the same as if I were to
calculate them manually." Isn't this what one would expect? Keeping in mind
that the nonbonded potentials in NAMD are all pairwise, all that is being
done under the hood is that NAMD is accumulating both the forces and
energies for those pairwise interactions involving atoms from group A with
atoms from group B. This is ultimately implemented in
ComputeNonbondedBase.h, where there is a loop that specifically acts on
atom pairs where one atom is in one of the interaction groups and the other
is in the other interaction group (unless pairInteractionSelf is set, in
which case it checks for both atoms being part of the interaction group).
That handles everything except for long range electrostatics, where there
are three separate PME grids calculated to allow the equivalent of the
three-group subtraction that Jeff mentioned (ComputePme.C).

On Mon, Oct 26, 2020 at 6:20 PM Mersch, Kacey <> wrote:

> I recently calculated the VDW Energies of one sidechain with many other
> sidechains individually on a static protein structure (no simulation was
> run) using the pair interaction calculation option. The values calculated
> are exactly the same as if I were to calculate them manually. With that in
> mind, what equation or process goes on when the pair interaction
> calculation is being performed?
> When performing the calculations manually, one would calculate the VDW
> energies in two sidechains and the interactions between them and then
> subtract away the self-interactions that occur between the atoms of each
> sidechain individually. This would leave you with only the VDW energies
> between the two sidechains.
> The manual states the outcome of using the pair interaction is the
> calculation of the energies between those two groups of flagged atoms. But,
> how is that calculated? is there a basic addition or subtraction process
> that occurs or something more complicated?
> initial config
> coordinates /VDW/17.pdb
> temperature 300K
> # output params
> outputname /VDW/
> binaryoutput yes
> # force field params
> structure /VDW/VDWDECOMP.psf
> parameters /VDW/par_all27_prot_lipid_na.inp
> paraTypeXplor off
> paraTypeCharmm on
> exclude scaled1-4
> 1-4scaling 1.0
> switching on
> switchdist 8.0
> cutoff 12.0
> pairlistdist 13.5
> stepspercycle 20
> pairInteraction on
> pairInteractionFile 17.pdb
> pairInteractionCol B
> pairInteractionGroup1 1
> pairInteractionGroup2 2
> One note for the question below. I am not asking about the calculations
> that occur when getting numbers for forces (VDW, ELEC). But, how those
> numbers are manipulated to get the energies between the two groups of
> atoms. So, my question is not about the LJ potential but, rather how does
> NAMD set aside the number calculated between the two groups of atoms.
> Any help is greatly appreciated!
> Kacey

This archive was generated by hypermail 2.1.6 : Thu Dec 31 2020 - 23:17:14 CST