Re: Impropers in Charmm and OPLS

From: johan strumpfer (
Date: Tue Jun 14 2011 - 10:01:13 CDT

Dear Henriette,

You are correct in your assumption. NAMD employ the CHARMM force field
using a harmonic function for impropers, and also AMBER/OPLS-style
force fields using a cosine function for impropers. It distinguishes
between the two types using the multiplicity column (the 6th column in
the IMPROPER definition). If the multiplicity is zero, then a harmonic
function is used, otherwise a cosine function is used.


Johan Strumpfer (
Theoretical and Computational Biophysics Group
3115 Beckman Institute
University of Illinois at Urbana-Champaign
405 N. Mathews
Urbana, IL 61801, USA

On Tue, Jun 14, 2011 at 9:33 AM, Henriette Elisabeth Autzen
<> wrote:
> Dear NAMD users,
> I am trying to simulate a protein in NAMD using the OPLS-AA FF. For this
> purpose, I have the Charmm formatted OPLS-parameters for proteins
> distributed at the Mackerell's homepage
> I have a question regarding how the impropers are treated in Charmm and
> The dihedral potential is modelled using a cosine function in both Charmm
> and OPLS so I would expect no problems when NAMD reads the dihedral
> parameters.
> However, the improper dihedrals are modelled differently using a harmonic
> potential in Charmm and a cosine function in OPLS.
> In the Charmm formatted OPLS protein file (par_opls_aa.inp) the impropers
> are written as if they are modelled by a cosine function, e.g. for the
> aromatic ring in phenyl alanine from the par_opls_aa.inp:
> H146 C145 C145 C145     1.1     2       180
> Can I assume that the OPLS parameters in Charmm format are treated according
> to the cosine function and not the harmonic function? If not, how is
> planarity enforced using OPLS in NAMD?
> Thank you in advance.
> Kind regards,
> Henriette

This archive was generated by hypermail 2.1.6 : Mon Dec 31 2012 - 23:20:26 CST