[gmx-users] Simulation of ordered molecular structures

[Peyman Yamin]

Dear gromacs users,

I'm trying to simulate amphiphilic molecules, building stable
structures by non-bonded interactions. I have seen in biological research, 
membranes and proteins therein are often subject of simulations. A cell 
membrane, I guess, could be an example of a structure built not by hard-core 
chemical bonds.

Let's say you know experimentally that molecule X, builds starlike
structures! Or a surfactant builds a sphere. Now to simulate this - to see 
e.g. the diffusion of other molecules in and out - does it make sense to 
build a sphere of, say 100 - the number you may know again from experiments - 
in a spherical geometry, and let them deal with eachother by a forcefield,
or it's necessary to explicitly fix their positions?  If yes, will it give
realistic results? In the end one could fix positions of 10, 100 and 1000 
molecules as a sphere, though only those with 100 are stable, in lab!

I would greatly appreciate any comment, from those who do membranes or
other alike predefined structures.

Thanks in advance,
Peyman