[gmx-users] Re: Mdrun interrupted using a triclinc box

[leafyoung81-group at yahoo.com]

--- Yinghong <xieyh at hkusua.hku.hk> wrote:

> Dear Yang Ye:
> 
> The PBC I used is one particular tpye of cubic box,
> which cross-section is a rhombus instead of
> rectangle.
> As we know, a hexagon can be divided into three
> rhombus, ok? We call it as unit cell, which is the
> box I employed. Normally, some distance is existing
> between solute molecule and box boundary in most
> simulations. But in my simulated box, the boudary is
> initially tangent to the solute, which means there
> is almost no distance between solute and box
> boundary. As you said, if your DNA exceed the
> boundary, one part will be in the right side, and
> the other will emergy in the left side. This is also
> the case I often encountered.
But this is quite okay because of PBC, atom is always
in contact with the nearest neighbours. 

> In my simulation, the solute is a nanotube, as the
> simulation, the tube will be outside the box and
> broken into several parts. And finally, the run will
> be stopped possibly due to the broken nanotube.
As far as I know that nanotube is a rigid structure,
do you use any contrainsts to hold the atoms together?
What kind of contraints?

In one of my project, I could only use SHAKE to fix a
8-atom system as a rigid system (LINCS could not do
so). But SHAKE is not good with PBC (pbc=xyz), which
makes molecule exploding and simulation crashes.

Hope this piece of information could help you.
> > but Gromacs told me that has no effect for unit
> cell representation.
> which will be told when I use "trjconv -ur tric -pbc
> whole (not inbox)" to define a tric box.  Do you
> know how to define a tric box as I mentioned before?
> 
I first estimate desired size of the box and then
change the last line of .gro file or 3rd line of .pdb
file.
> Thanks for your help and discussion.
> 
> Xie Yinghong
> Hong Kong University.
> 
> 
> > Dear users:
> >
> > In my system, I chose a triclinc box (monoclinic
> cell) as my periodic
> > boundary condition, which is defined as follows:
> >
> > editconf -f *.pdb -o -bt tric -box 5 5 10 -angles
> 90 90 120 -c
> > genbox -cp out -cs -p topol -o b4em
> > grompp -v -f em -c b4em -p topol -o box
> > trjconv -f b4em -s box -o b4em2.gro -ur tric -pbc
> inbox
> >
> > If I used "-pbc inbox", my simulated molecule
> would be broken.
> >
> What do you mean by "broken"? In my simulation,
> molecule will gradually
> slide to the boundary of the box and even move out
> from it. But at long
> as it is still kept at a good shape, there is no
> impact to MD
> simulation. For DNA, there could be case that one
> strand on the left
> boundary and another on the right boundary. A awk
> script or program will
> do the trick and make analysis proceeding.
> 
> > And, if "-pbc whole" was adapted, the molecule
> could be kept intact,
> > but Gromacs told me that has no effect for unit
> cell representation.
> >
> where did this "gromacs told me..." come from?
> 
> > In my simulations, I empolyed "-pbc whole" in
> order to keep molecule
> > intact, but after ~200ps simulation, mdrun is
> interrupted, and the
> > initial intact molecule is also broken at this
> moment.
> >
> Also, what do you mean "broken" here?
> 
> > So, Could you tell me that the way I defined a
> triclinc box *(not
> > hexagonal box)* is right?
> >
> > Which option should I choose in "trjconv" for -ur
> and -pbc?
> >
> > What made my simulation stopped?
> >
> > Because this is my first time using triclinc box,
> many things are very
> > strange for me.
> >
> > Any help is appreciated.
> >
> > Xie Yinghong
> >
> > Hong Kong University
> >
>
>------------------------------------------------------------------------
> >
> >_______________________________________________
> >gmx-users mailing list
> >gmx-users at gromacs.org
> >http://www.gromacs.org/mailman/listinfo/gmx-users
> >Please don't post (un)subscribe requests to the
> list. Use the 
> >www interface or send it to gmx-users-request at
> gromacs.org.
> >
>