AMBER Archive (2008)

Subject: RE: AMBER: Electrostatic Energy Components

From: Ross Walker (ross_at_rosswalker.co.uk)
Date: Fri Nov 21 2008 - 15:36:20 CST


Hi Carlo,

> I would like to ask a question that probably you will
> consider trivial. What is the difference between the
> energy components 1-4 EEL and EELEC ?

I don't consider this a trivial question in fact I think no MD paper makes a
good job of actually explaining how the force field works. For example most
people write the AMBER or CHARMM force fields incorrectly in that they
include the sum over VDW and electrostatics to be over all atoms within a
cut off. However, they forget that this summation actually has an exclusion
list associated with it.

Essentially VDW and EEL interactions are NOT calculated between 1-2 (bonds)
and 1-3 (angles) atoms that are covalently bonded. The reason for this is
that these interactions are effectively parameterized into the bond and
angle terms. A similar situation exists with the dihedral interactions
except here the complete VDW and electrostatic interactions are not included
within the parameterization of the dihedral term. However, calculating the
full electrostatic and VDW interaction for 1-4 bonded atoms can make it
difficult to obtain the desired behavior and so in the current AMBER force
fields (GAFF, FF99SB, FF03 etc) the 1-4 interactions are calculated but then
scaled by scee for electrostatics and scnb for VDW.

Hence everything that is not bonded or separated by more than 3 bonds is
included in the full electrostatic and VDW interaction as you see in the
force field equation. All 1-2 and 1-3 interactions are skipped and all 1-4
interactions are divided by 1.2 for EEL and 2.0 for VDW. (Note these values
are a function of the way the force field was parameterized and so should
never be changed - for example GLYCAM06 does no scaling of 1-4 terms and so
these values are set to 1.0).

Hence in order to do this the code first builds a direct space list of pair
wise interactions and then subtracts from this list all 1-2, 1-3 and 1-4
pairs. It then calculates the EEL and VDW terms which is what you see in the
output as EELEC and VDWAALS (note in periodic simulations the PME EEL also
ends up in EELEC). Then when doing the dihedral calculations for each unique
1-4 term it also calculates EEL and VDW and scales the energy and forces
appropriately. This energy could then simply be added to the EELEC and
VDWAALS terms but for troubleshooting etc it is instead reported separately
as 1-4 EEL.

> Why is
> the EEL 1-4 term usually positive while EELEC is
> generally negative ?

I think that the 1-4 term being generally positive is a function of the
typical distribution of charges around atoms separated by 3 bonds, there is
certainly no requirement for it to be positive. Note though that the actual
interaction between such atoms is really a function of both the 1-4 terms
and the dihedral terms.
 
> I also noticed that if I add together EEL 1-4 and EELEC,
> this correlation with EGB becomes weaker. Why is it so?

I'm not sure about this and others might be able to comment more on this but
the main point is that the force field here isn't strictly pairwise
decomposable, I.e. the 1-4 term also has a component to it that is
parameterized into the dihedral term.

All the best
Ross

/\
\/
|\oss Walker

| Assistant Research Professor |
| San Diego Supercomputer Center |
| Tel: +1 858 822 0854 | EMail:- ross_at_rosswalker.co.uk |
| http://www.rosswalker.co.uk | PGP Key available on request |

Note: Electronic Mail is not secure, has no guarantee of delivery, may not
be read every day, and should not be used for urgent or sensitive issues.

-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber_at_scripps.edu
To unsubscribe, send "unsubscribe amber" (in the *body* of the email)
      to majordomo_at_scripps.edu