Re: [AMBER] cis form: avoiding atom overlap in all-atom modeling

From: Aron Broom <>
Date: Thu, 18 Oct 2012 12:53:40 -0400

The thing is that even if the H-atoms would have VDW overlap, that is only
for the completely cis case. A shift around that dihedral of only a few
degrees would alleviate that problem, but would still be overall cis.

You can use methods such as umbrella sampling (supported with pmemd and
pmemd.cuda) or metadynamics (I think currently only on sander) to get a
better idea of the energy landscape around that bond (rather than relying
on chance observations during your equilibrium MD run).

If the cis-form is really more stable, and you are not seeing that by MD,
it suggests a problem with parameterization of the forcefield. There are a
lot of ring structures there, and I don't know if GAFF or AMBER are
particularly well setup to handle accounting for the effects of
pi-electrons (maybe someone can comment). Moreover, given that your ring
hydrogens should have a fairly high partial positive charge (owing to the
electron withdrawing nature of the ring), there may also be significant
polarization effects that you aren't taking into account unless you use a
polarizable forcefield.

Having said all that, maybe it can still work but you just haven't derived
the charges accurately. How did you get the partial charges, did you use
the RESP server?


On Thu, Oct 18, 2012 at 12:42 PM, bio lab <> wrote:

> Dear Amber community members,
> I would like to know your opinion whether it is possible or not to observe
> a trans-to-cis transition of Resveratrol in MD simulation by using the
> GAFF/Amber force-field. This molecule is present in nature in both forms
> (and the cis form is also more stable), as it can be seen on the following
> Wikipedia image:
> In the cis form the two 6-atom rings are very close. It seems to me that
> the H atoms that are attached to the two closest C atoms would overlap when
> modeled with an all-atom force-field like Amber, thus inhibiting the
> possibility of obtaining the cis form. Indeed, in MD simulation I observe
> many (short-lived and incomplete) rotations of the corresponding dihedral
> angle: they are sufficiently large to overcome the energy barrier (<90),
> but the molecule always returns back to the trans form (180) and never
> ends up in the cis one (0).
> Therefore, my question is the following: is it possible to observe this in
> MD simulation? Or does it require any kind of special modeling of the
> molecule, for instance the two 6-atom rings cannot be considered planar in
> the cis form? Or should I try in this case a united-atom force-field?
> Thanks in advance for any comment or suggestion.
> _______________________________________________
> AMBER mailing list

Aron Broom M.Sc
PhD Student
Department of Chemistry
University of Waterloo
AMBER mailing list
Received on Thu Oct 18 2012 - 10:00:05 PDT
Custom Search