[gmx-users] the sign of the bonding energy term

[Christoph Freudenberger]

Yong Kong wrote:
> Dear GROMACS team/users!
> 
> I am a new uesr, and have a general question on  the bonding interactions in
> GROMACS (bonds, angle and dihedral). It seems that GROMACS treats all the three
> types of bonding as positive terms. That means for a non-protein system with
> zero partial charge on atoms the calculated potential energy must be positive,
> is my understanding right? 
Well, if you neglegt all non-bonding terms of the energy (LJ and Coulomb) and have
all bonds, angles and dihedr. in their equilibrium positions the pot. energy
will be zero. But it will not be negative, thats right.

> And if that's true, the energy minimization still
> works? 
AFAIK both sd and cg look for minima on the PES. I don't think it matters
for the algorithm whether the value at the minimum is positive or negative.

> Should I change the signs of the bonding parameters manually to maintain
> the total/potential energy negative?
No you shouldn't. The 'real' potetial energy of molecule is negative
because the electrons are moving within the electrostatic field of the
nuclei. If I understand things right, these terms are not treated in
MM/MD forcefields. The bonding terms give the potential energy of
the molecule relative to it's QM ground state. Of course it's pot. energy
cannot become smaller than that.

hope that helps
-- 
Christoph Freudenberger
Abt. Organische Chemie I AK Siehl - Uni Ulm -Tel: ++49-731-502-2785