From: Marcos Sotomayor (sotomayo_at_ks.uiuc.edu)
Date: Thu Jul 31 2008 - 13:01:12 CDT
There are several advantages in using a water box and periodic boundary
conditions (PBC) in SMD simulations. As Axel already mentioned,
with PBC you get a more realistic situation: your protein is always
surrounded by water, there are no artifacts caused by water/vacuum
boundaries, you can control pressure in your system, and you can use PME
to get more accurate electrostatics.
Most of the current SMD simulations are done using PBC with a water box
that is large enough to accommodate the stretched/unfolded protein.
Sometimes this means having a very large water box! If you don't have
enough computer power, such a large water box might be prohibitive
(that's why the first SMD simulations were done in water spheres....). So,
in order to avoid interactions with the periodic image, just enlarge your
water box in the pulling direction as much as is needed to fit the
As for your "pushing" SMD simulation, you may want to apply the forces to
the two termini carefully. As far as I remember, the regular SMD will
apply forces only along the pulling/pushing axis. If you implement
something similar, your peptide will likely rotate. It shouldn't be very
difficult though to apply moving harmonic constraints in all three
dimensions using the NAMD Tcl forces interface.
Hope that helps,
On Thu, 31 Jul 2008, Roman Petrenko wrote:
> Dear Axel,
> thank you for your thoughts about water, but as to lammps i am not
> willing to learn another simulation package just because of some extra
> SMD functionality (unless i get 10 more reasons why lammps is better
> than namd :) )
> Back to namd and SMD in water with periodic boundary conditions. Is
> there any manual online on how to do it correctly? Namely, how the SMD
> atom is not going to hit the fixed atom (at the other end of the
> chain) once it goes outside of the box and it's image appears on the
> other side?
> On Thu, Jul 31, 2008 at 10:16 AM, Axel Kohlmeyer
> <akohlmey_at_cmm.chem.upenn.edu> wrote:
>> On Thu, 31 Jul 2008, Roman Petrenko wrote:
>> RP> A general question about SMD and water.
>> RP> I've noticed in many articles that a protein is solvated and then SMD
>> RP> simulations are used (one atom fixed, another is pulled) without any
>> RP> boundary constraints.
>> RP> Is there any specific reason besides having significantly smaller
>> RP> number of water molecules?
>> RP> I want to simulate a small peptide with SMD (so the number of waters
>> the number of water molecules is always important. we always use
>> far too few waters (by orders of magnitude!) to not see any impact
>> of the number of waters.
>> RP> is not that important) and to use either cylindrical or spherical
>> RP> boundary conditions. The N- and C-terminal CA atoms are supposed to be
>> i would say that is running a very high risk of having your results
>> being dependend on the position of the peptide in the water area (you
>> have surface dipoles!). why not use periodic boundaries? you can use a
>> smaller box and the position of your peptide relative to the center of
>> the water system does not matter at all.
>> RP> pulled towards each other along parallel trajectories.
>> RP> Nterm------(force)------------------->
>> RP> \ ------
>> RP> --- / \ (peptide)
>> RP> \ / ---------
>> RP> ------- \
>> RP> <------------(force)-------------Cterm
>> RP> Does this design have an advantage that the whole peptide is always
>> RP> going to be at the center of either water sphere or water cylinder?
>> please note that this is not quite the same experiment than
>> pushing/pulling only one end. the pulling speed is usually
>> very high, so that the resulting free energy profile is not
>> the same. the area where you apply the pulling force is usually
>> behaving non-physical.
>> RP> The reason i ask is because maybe someone have already done this and
>> RP> ruled out this design as being not effective.
>> i cannot tell which option would be more effective, but you could
>> try it out with the SMD module for LAMMPS, that our group is working
>> on and that offers this functionality (and other enhancements).
>> Axel Kohlmeyer akohlmey_at_cmm.chem.upenn.edu http://www.cmm.upenn.edu
>> Center for Molecular Modeling -- University of Pennsylvania
>> Department of Chemistry, 231 S.34th Street, Philadelphia, PA 19104-6323
>> tel: 1-215-898-1582, fax: 1-215-573-6233, office-tel: 1-215-898-5425
>> If you make something idiot-proof, the universe creates a better idiot.
> Roman Petrenko.
> Physics Department
> University of Cincinnati
This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:48:07 CST