Hi,
After running several tests I think I may know what is going on here.
Were you calculating residue interaction energy within a single
molecule? When I do pairwise decomposition with sander, idecomp 3
results are the same as idecomp 4 results if I'm only looking at
residue pairwise interactions for a single molecule. What's more, it
seems that 1-4 van der Waals and electrostatics are added to the total
van der Waals and electrostatics no matter what. Here is an example of
a single molecule with two residues:
SANDER IDECOMP3:
resid1 ->resid2 |internal |vdw |eel |pol |sas
======================================================================
TDC 1-> 1 0.000 0.164 -19.266 -89.941 0.000
TDC 1-> 2 0.000 0.089 -6.803 -0.741 0.000
TDC 2-> 1 0.000 0.089 -6.803 -0.741 0.000
TDC 2-> 2 0.000 2.073 27.719 -16.849 0.000
CPPTRAJ (using a development command that calculates energy, but the
vdW/elec results are identical to 'pairwise'):
#Frame dEvdw141.2 dEelec141.2 dEvdw1.2 dEelec1.2 ide4_V1.2 ide4_E1.2
1 0.567 3.741 -0.478 -10.543 0.089 -6.803
The columns after #Frame are delta 1-4 vdW energy, delta 1-4 elec
energy, delta vdW energy, delta elec energy, sum of delta 1-4 vdW and
delta vdW energies, and sum of delta 1-4 elec + delta elec energies.
To match sander where the energies are split between residues, the
delta here is being calculated as:
(ERES1+2 - ERES1 - ERES2) / 2.0
You can see that the sander results match the cpptraj results only if
the 1-4 terms are included. However, if I look at two separate
molecules, each 1 residue:
SANDER IDECOMP3:
resid1 ->resid2 |internal |vdw |eel |pol |sas
======================================================================
TDC 1-> 1 0.000 0.164 -19.266 -95.201 0.000
TDC 1-> 2 0.000 -0.001 11.008 -10.868 0.000
TDC 2-> 1 0.000 -0.001 11.008 -10.868 0.000
TDC 2-> 2 0.000 3.874 61.173 -83.685 0.000
CPPTRAJ:
#Frame dEvdw1.2 dEelec1.2
1 -0.001 11.008
Now the sander vdw/elec results match cpptraj exactly. The bottom line
is that energy decomposition code in sander has a lot of bookkeeping
that it does under the hood, and you have to be careful when comparing
those results to the simple straightforward energy calculation in the
cpptraj 'pairwise' command.
Hopefully this clears things up a bit,
-Dan
PS - Version 14.17 of cpptraj is now available and includes many
important fixes, so I recommend updating.
--
-------------------------
Daniel R. Roe, PhD
Department of Medicinal Chemistry
University of Utah
30 South 2000 East, Room 307
Salt Lake City, UT 84112-5820
http://home.chpc.utah.edu/~cheatham/
(801) 587-9652
(801) 585-6208 (Fax)
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Thu Nov 20 2014 - 10:30:02 PST
