Re: Charmm 1,4 Interactions - Scaling Question

From: Jan Saam (
Date: Thu Jun 16 2005 - 18:20:22 CDT

Dear Paul,

A scaling factor of 1.0 means that the 1,4 interactions are fully
included. I've once spoken to Alek Mackerell about this issue and he
emphasized the importance of this scaling factor in order to obtain
correct results.

Neveretheless the CHARMM force field specifies modified VDW interactions
for some atom types. You can find these in the parameter file. Here is
an example from the NONBONDED section:

!V(L-J) = Eps(i,j)*[(Rmin(i,j)/r(i,j))**12 - 2(Rmin(i,j)/r(i,j))**6]
!epsilon [kcal/mole]: Eps(i,j) = sqrt(eps(i) * eps(j))
!Rmin/2 [A]: Rmin(i,j) = (Rmin(i) + Rmin(j)) / 2
!atom ignored epsilon Rmin/2 ignored eps,1-4 Rmin/2,1-4

C 0.000000 -0.110000 2.000000 ! ALLOW PEP POL ARO

CT1 0.000000 -0.020000 2.275000 0.000000 -0.010000 1.900000

For atom type C only the parameters epsilon and Rmin/2 are specified
thus no scaling is applied for 1-4 interaction (they are fully
included). For atom CT1 though, the additional parameters eps,1-4 and
Rmin/2,1-4 are given. These values are used for 1-4 interactions.

The problem with 1-4 interactions is that their nonbonded contribution
is already partially included in the corresponding dihedral parameter.
Scaling down the 1-4 VDW interaction can help avoiding double accounting
is a means of fine tuning the force field. You actually only have to
worry about it if you are adding new parameters to the force field.

Hoping this helps,


Paul R Brenner wrote:
> I have been reading the recent J Comput Chem paper by Dr. Mackerell
> (DOI 10.1002/jcc.20082) and am now questioning the mechanism by which
> 1,4 interactions are included/excluded for proteins in NAMD.
> Dr. Mackerell states: "1,4 Nonbonded interactions, involving atoms
> separated by three covalent bonds, do have nonbonded contributions as
> well as their spatial relationship being influenced by the internal
> terms, including the dihedral term. However, different biomolecular
> force fields treat the 1,4 nonbond interactions differently... The
> all-atom CHARMM biomolecular force fields do not scale the 1,4 terms
> (scale factor = 1).
> With a scale factor of 1, are 1,4 LJ contributions fully included or
> excluded in NAMD?
>> From the NAMD Tutorial:
> When the scaled1-4 exclusion policy is used (as it should with the
> CHARMM force field) nonbonded interactions of atoms separated by
> three bonds (i.e., atoms 1 and 4 in the chain 1-2-3-4) are modifed.
> Even if the scaling factor for electrostatics is 1.0 (as it should be
> for modern CHARMM force fields), special modified van der Waals
> parameters are used for 1-4 pairs of atoms for which they are
> specified
> So how does 1,4 scaling work when there is a special set of 1,4 LJ
> parameters specified? Will the special parameters be recognized if
> scaled1-4 is not specified.
> Any insight on the standard/best method for specifying the 1,4
> interactions in NAMD using the CHARMM force field, would be
> appreciated.
> Thanks,
> Paul R Brenner, P.E. Graduate Research Assistant Computer Science and
> Engineering The University of Notre Dame

Jan Saam
Institute of Biochemistry
Charite Berlin
Monbijoustr. 2
10117 Berlin
+49 30 450-528-446

This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:40:51 CST