[gmx-users] Gromacs Tutorial by Justin

Justin Lemkul jalemkul at vt.edu
Sat Mar 10 20:56:55 CET 2018

On 3/10/18 1:24 PM, Faezeh Pousaneh wrote:
> Dear Justin
> One more question please,
> Do we need to have constraint in this model of CO2?  Isn't it so that we
> have fixed the bond lengths and angles (180) in the topology by this method?

The molecule behaves as a rigid rotor. The math is all laid out in the 
tutorial. Since the construction of the O virtual site positions depend 
on the bond length connecting the two mass centers (expressed as a 
fraction), this (1) makes the math easier, (2) aligns well with known 
and easily demonstrated physics, and (3) prevents the need for 
parametrizing the bond connecting the two mass centers.

> I run 2 simulations, one with constraints = all-angles  and the other
> with constraints
> = all-bonds, I get different results from them. Why?

Please define "different results."

Every simulation will produce different results; no two trajectories 
will be binary identical, but their properties should agree within 
error. I see no reason why there should be any difference here. There 
are no angles to be constrained, nor are there any bonds to be converted 
to constraints. There is one explicit constraint in the topology. So 
that keyword should be having no effect whatsoever.



Justin A. Lemkul, Ph.D.
Assistant Professor
Virginia Tech Department of Biochemistry

303 Engel Hall
340 West Campus Dr.
Blacksburg, VA 24061

jalemkul at vt.edu | (540) 231-3129


More information about the gromacs.org_gmx-users mailing list