[gmx-users] Dihedral Question

[Justin A. Lemkul]

Fabian Casteblanco wrote:
> Hello all,
> 
> I'm having trouble finding information that explains the relationship
> between 1,4 pair interactions and dihedrals.  I'm building a drug
> molecule using CGenFF, and extension of CHARMM27 and since the drug is
> a bit complex, it has several dihedral parameters that are missing.  I
> ran them with the missing parameters and then I reran it with
> estimated parameters based on other dihedral parameters for similar
> dihedrals and they look somewhat similar but still have noticeable
> differences.
> 
>>From what I understood, the 1,4 pair interactions should already be
> setting most of the dihedral angles up and the dihedrals are simply
> added energy consequences inorder to keep the specified torsion angle.
>  Is this the correct way 1,4 pair interactions and dihedrals work
> together?  Also, if I were to simply run without using the dihedral
> parameters, would it make a huge difference when extracting
> themodynamics information from a mixture of this single drug molecule
> in different types of solvent?
> 

1-4 interactions and dihedrals work in concert, but it is not true that 1-4 
interactions "set up" the dihedrals.  1-4 interactions are short-range 
intramolecular nonbonded terms, and dihedrals are bonded terms.  Different force 
fields treat these interactions in different ways.  Some force fields rely scale 
1-4 interactions by certain factors (FudgeLJ and FudgeQQ in the forcefield.itp 
file), while others do not apply any scaling.  Thus there is a balance between 
the ability of the dihedral to move through space and the repulsive forces 
between the atoms.

The effect on hydration free energies, etc may or may not be small.  If a 
particular missing interaction affects the conformation of your molecule, it may 
in turn influence the way water molecules (or whatever solvent) assemble around 
it.  If you don't have correct dihedrals and do not sample correct 
configurations, you may not get the right result.  It's hard to predict the 
outcome, but this is one reason why developing good parameters for arbitrary 
molecules is so difficult.

-Justin

-- 
========================================

Justin A. Lemkul
Ph.D. Candidate
ICTAS Doctoral Scholar
MILES-IGERT Trainee
Department of Biochemistry
Virginia Tech
Blacksburg, VA
jalemkul[at]vt.edu | (540) 231-9080
http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin

========================================