Re: harmonic-restraints/colvars

From: Francesco Pietra (
Date: Tue Jan 14 2020 - 09:42:43 CST

Hi Jerome
Thanks, in particular for illustrating the drawbacks of restraints/colvars
side by side.

Along the various attempts, I also used a homogeneous set of harmonic
restraints, playing on the values at column B for different atoms. However,
in any case, by decreasing the constraintScaling below 10, the coordination
between the small organic molecule at its nucleoside ligands at the binding
pocked was rapidly deteriorated with respect to the X-ray diffraction data.
I mean that hydrogen bonds (already weak from Xray diffraction patterns)
and pi-stackings were completely broken.

Probably getting a MD stable such system in TIP3 without substantial
constraints is very difficult to attain, if at all. Now, I can not think
better than trying with distance colvars at the binding pocket added to the
harmonic restraints (colum B).


On Tue, Jan 14, 2020 at 11:25 AM Jérôme Hénin <> wrote:

> Hi Francesco,
> That is allowed at a technical level: NAMD's harmonic restraints and
> Colvars will work side by side - I suppose you know that, but I am
> repeating it for the benefit of others who'd read this thread based on its
> subject line.
> I cannot tell you what is scientifically relevant for your specific
> system: any set of restraints is legitimate, but then the simulations
> explore a different ensemble, and that of course limits the information
> that can be gained from them. Possibly to the point where everything
> significant is restrained, and the simulation becomes completely
> uninformative.
> Jerome
> On Tue, 14 Jan 2020 at 10:49, Francesco Pietra <>
> wrote:
>> Hello: is that allowed using a mixture (for different groups of atoms) of
>> harmonic restraints (by creating a pdb file marked with different values on
>> column B) and distance colvars?
>> Could that be equivalent to, or be more appropriate than by, using a
>> mixture of distance and rmsd colvars for the same different groups of atoms?
>> The question is related to getting a MD stable system by keeping an
>> organic ligand at its binding pocket in a model that results from cutting
>> the natural huge biomolecule all around the ligand.
>> So far (despite parameterization of the ligand at the state of the art
>> with CGenFF, whereby it proved stable in TIP3P for hundreds of ns) I was
>> unable to get a stable system when removing all restraints/colvars except
>> at the region where the biopolymer was cut..
>> thanks for a general advice
>> francesco pietra

This archive was generated by hypermail 2.1.6 : Thu Dec 31 2020 - 23:17:12 CST