Re: MD simulation of Linear Molecules (like 2-butyne£©

From: Jim Phillips (
Date: Sat Dec 19 2015 - 21:53:01 CST

I don't see an alternative to your proposed long-range dihedrals, and I
can't think why they wouldn't work as expected. You'll probably want to
add those via a patch after autogenerating the rest. Make sure the
dihedral parameters assigned to 6-8-9-10 are zero as it is degenerate.

I think linear angles are fine, as I recall people simulating CO2.


On Sat, 19 Dec 2015, Ō¬Š”¾§ wrote:


I'm having trouble with the MD simulation of a molecule with linear part (structure of it:, and I'm not sure in what way should I treat the linear angles and the related dihedral angles in MD after I finished the parameterization of the molecule.

These are what I tried:
Since it seems that CHARMM force field uses redundant internal coordinate similar to Gaussain 09, I am trying to edit the psf file part involving the linear angles as in Gaussian ( C. Peng, P. Y. Ayala, H. B. Schlegel, and M. J. Frisch, ”°Using redundant internal coordinates to optimize equilibrium geometries and transition states,”± J. Comp. Chem., 17 (1996) 49-56. ).

I deleted the dihedrals involving the linear bends and added the long range dihedrals (D 7,6,10,11; D 7,6,10,15; D 5,6,10,11; D 5,6,10,15) like in Gaussian. I'm not sure if it's ok to manually edit the psf file, but what else can I do to define the long range dihedral angles in NAMD? Or there just should not exist such long range dihedral angles in MD simulation?

As to the linear angles (L 6,8,9; L 8,9,10), Gaussian needs 2 orthogonal linear angle bends to define 1 linear angle, in this case L (6,8,9,5,-1) L(6,8,9,5-2) L (8,9,10,15,-1) L(8,9,10,15,-2). Is there anything I can do to define the linear angles in NAMD instead of just treating it as an angle with equilibrium value equal to 180 degree?


This archive was generated by hypermail 2.1.6 : Thu Dec 31 2015 - 23:22:21 CST