Dear LAMMPS users,
I set up an isolated vesicle and am going to do some CGMD simulations by using a solvent-free CG model. For the solvent-free model, it is hard to apply inner pressure by water.
I tried to modify the “fix indent” to apply a radial force to all the atoms in the cell. But due to the cell configuration is flexible and not a sphere always, I cannot make sure the radial forces keep vertical to the cell surface. The radial force doesn’t seem to be consistent with what should be in the real case.
Now I am trying to improve my modified “fix indent” command. For every atom on the cell, I use the coordinate information of all the neighbors to calculate the tangent plane. Then we can get the “pressure” by the radial force resolved on the nomal of the tangent plane. But it is much more complicated.
Do you have any better suggestions? Can I get the same effect by using existing commands in LAMMPS? Any suggestion is highly appreciated.
Best,
Wenpeng
I’m not clear conceptually on what you want to do.
What does “apply inner pressure by water” mean?
Steve
Dear Steve,
Thank you for your attention.
For the large cell size (1 micron) I am modeling, I need to choose a most coarse grained model for cell membrane. Solvent-free coarse grained model is used in my simulations for lipids of cell membrane. Since it is a solvent-free model, water molecules are implicit. It is impossible to use water molecules to apply inner pressure on cell. However, osmotic pressure of cell is an essential parameter I need to consider.
I am wondering if there is an easy way to apply inner pressure by known coordinates or velocities of every atom on the cell membrane. The thickness of the cell membrane is made up of 6 atoms.
Now I wrote a fix command modified by the "fix indent ". For every atom on the cell, I used all its neighbors to calculate the tangent plane. Then I can apply the “pressure” on every atom along its normal of the tangent plane. But it is very complicated, and sometimes disabled when the cell deforms largely.
Do you have any better suggestion? I hope my explanation is clear to you.
Best,
Wenpeng Zhu
Dear Steve,
Thank you for your attention.
For the large cell size (1 micron) I am modeling, I need to choose a most
coarse grained model for cell membrane. Solvent-free coarse grained model is
used in my simulations for lipids of cell membrane. Since it is a
solvent-free model, water molecules are implicit. It is impossible to use
water molecules to apply inner pressure on cell. However, osmotic pressure
of cell is an essential parameter I need to consider.
I am wondering if there is an easy way to apply inner pressure by known
coordinates or velocities of every atom on the cell membrane. The thickness
of the cell membrane is made up of 6 atoms.
Now I wrote a fix command modified by the "fix indent ". For every atom on
the cell, I used all its neighbors to calculate the tangent plane. Then I
can apply the "pressure" on every atom along its normal of the tangent
plane. But it is very complicated, and sometimes disabled when the cell
deforms largely.
Do you have any better suggestion? I hope my explanation is clear to you.
not clear, if this is a choice for you, but have a look at the "fix
oneway" command.
axel.
Hmm. 1um is pretty large indeed.
I haven't thought much about this, so I might embarrass myself with this reply:
Would it be feasible to put a dilute gas of repulsive particles inside
the vesicle? (These particles do nothing but push on the membrane.
You could turn off interactions with the other particles using the
"neigh_modify exclude" command.
http://lammps.sandia.gov/doc/neigh_modify.html)
In other words, create a very dilute gas which interact repulsively
with the lipids [and each other], and then put them inside the
vesicle.
I would make the particles very large, to compensate for using small
number of them. (To keep the pressure high, decrease the
effective/available volume, by making the particles large.)
I THINK the way to do this efficiently is to do that would be to use a
combination of "pair_style colloid", "neighbor multi", and
"communicate multi". (I'm not sure if "pair_style colloid" is
necessary to make it efficient, but try that first.)
http://lammps.sandia.gov/doc/pair_colloid.html
I'm curious. Can you tell us what lipid model you are using?
(I'm wondering if I should add it to moltemplate.)
Cheers!
Andrew
P.S.
I don't think this is what you were looking for, but also check out
fix spring/rg
http://lammps.sandia.gov/doc/fix_spring_rg.html
and, if you have not discovered this yet, check out:
http://moltemplate.org
Dear Andrew,
Thank you very much for your reply.
I think dilute gas is a good idea for my case. It can provide cell inner pressure with relatively low computational cost. It is worth trying.
The CG lipid model I use is a solvent-free model. It is developed by Cooke et al. (J. Chem. Phys. 123, (2005)). The link of the paper is shown as below:
http://scitation.aip.org/content/aip/journal/jcp/123/22/10.1063/1.2135785
which I found in a review paper of solvent-free lipid model (http://link.springer.com/article/10.1007%2Fs00249-005-0013-y).
A Nature paper (Reynwar et al. Nature. (2007)) and a Nature Nanotechnology paper (Shi et al. Nature Nanotechology. (2011)) use this solvent-free CG lipid model to investigate the endocytosis of capsid and nanomaterials. They also provide some information of this CG model in their supplementary materials. I think this lipid model is convincing and good for large-scale simulations.
Best,
Wenpeng Zhu