[gmx-users] Creating a monolayer from Martini bilayer

[Mirco Wahab]

On 03.06.2013 18:43, Neha wrote:
> I am a new user to Gromacs and am working on lipid simulations with the
> Martini forcefield. The Martini website provides a pre-equilibrated DPPC
> bilayer, and I was wondering if there were any told that would allow me to
> convert this bilayer into a monolayer. For periodic boundary conditions to
> work, I was thinking of a stack that would go like
>
> Water
> Lipid Heads
> Tails
> Vacuum
> Tails
> Lipid Heads
> Water

Is this the 128 DPPC system from
http://md.chem.rug.nl/cgmartini/index.php/downloads/example-applications/66-dppc-membrane
(2x 64 DPPC + 2000 W)?

> Is there any way to split apart the provided bilayer using some combination
> of Gromacs tools and introduce a space of at least 10 nm which is apparently
> the distance needed for the tails to not interact with each other?

The layers are, 'afaics', somehow interdigitated and placed in a box
normal to z with a z height of 10nm and centered at about 5nm. They
are not sorted (first 64 lower, secod 64 upper layer).

> If there is no way of simply splitting the bilayer, what would you recommend
> for creating a simulation of lipid monolayers from a single DPPC molecule? I
> feel like I could use genconf but it might require too much equilibration. I
> am hopeful that there is some way of simply working with the provided
> bilayer.

What I would do: write a simple script, read 12 lines (one molecule) and
check for the z coordinate of the NC3 type (first entry of each 12 line
record). Then, decide:
  - if NC3[z] > 6 then add 1.0 to all 12 z coordinates (nm)
  - if NC3[z] < 4 then keep it
write the 12 lines back to another file, either modified or unmodified.
next 12 lines/molecule (128 times total)

Then, the waters. This is somehow arbitrary. Maybe you could check if
each water's (W) z coordinate is below 5, so keep it - or above 5,
so move it up. This will, of course, require some steepest descent
minimization afterwards.