namd energy

From: Leandro Martínez (leandromartinez98_at_gmail.com)
Date: Tue Feb 05 2008 - 03:39:30 CST

I'm trying to analyze some non-bonded energy terms with NAMD energy.
I don't understand exactly the cause of some differences in the outputs.
I'm computing the non-bonded interaction energies between two atoms,
in one frame, for instance

selection 1: index 1455
selection 2: index 44

The atoms, in this frame (the only one in my tests), are 2.2 angstroms
appart.

If use the NAMD energ plugin to compute their non-bonded interaction
energies,
I get different results, depending on the cutoff I use. For example:

*
cutoff 5 angs:*
Frame Time Elec VdW Nonbond Total
0 0 -183.093 +21.1309 -161.962 -
161.962
*
cutoff 10 angs:*
Frame Time Elec VdW Nonbond Total
0 0 -236.93 +21.1309 -215.799 -
215.799
*
cutoff 15 angs:*
Frame Time Elec VdW Nonbond Total
0 0 -247.661 +21.1309 -226.53 -
226.53
*
cutoff: 100 Angs *
Frame Time Elec VdW Nonbond Total
0 0 -260.94 +21.1309 -239.809 -
239.809

The switch distance is always equal to the cutoff.

If I try to compute the energy myself, using the distance and the
force-field
parameters, I get:

Elec: -256.39
VdW: +21.11
Total: -235.28

So I have two problems: one is that the electrostatic energy changes if
one changes the cutoff in the range 5 to 100 angs, even while the
atoms are 2.2 angs appart. The other is that I don't get exactly the same
value for the VdW energy if I do the calculation myself, although these
values are coherent to each other in the NAMD calculations.
The system has periodic boundary conditions of about 80 angs, so I don't
see how by changing the cutoff between 5 and 10 angs the result changes.

Any help will be appreciated,
Leandro.

This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:49:13 CST