[gmx-users] fatty acid monolayer - surface pressure isotherm

McMullen, Roger L RMcMullen at ispcorp.com
Fri May 23 01:01:02 CEST 2003

Thanks Pedro.

1) I am currently increasing the simulation length to see if that helps.  In
some of the earlier publications (early 1990's), they were performing rather
short simulations (on the order of 400 ps) and were surprisingly getting
pretty good isotherm data.
2) I have pressure coupling turned to off so there is no setting for tau-p.
We wanted to keep the box size constant and monitor the pressure.  Is our
thinking about this wrong?
3) We have plans to increase the number of molecules in the system, but
thought that at present we should get reasonable data with 64 molecules
since previously published results using the same number of fatty acids
resulted in an experimentally representative Pi-A isotherm.
4) We are using the ffgmx forcefield.

Roger McMullen

-----Original Message-----
From: Pedro Alexandre Lapido Loureiro [mailto:paloureiro at biof.ufrj.br] 
Sent: Thursday, May 22, 2003 3:32 PM
To: gmx-users at gromacs.org
Subject: Re: [gmx-users] fatty acid monolayer - surface pressure isotherm

> Greetings all.
> I have a fatty acid monolayer system consisting of 64 arachidic acid
> molecules in vacuum.  We are trying to simulate a Langmuir trough
> experiment
> in which the lateral pressure is monitored as a function of the packing
> density.  The x and y vectors of the box are initially set at values that
> provide a monolayer packing density of 18.5 A^2/molecule.  
> At this packing density, I first perform a MD simulation for 1000 ps with
> position restraints on all atoms.  This is followed by another 1000 ps
> simulation in which the acid headgroups are position restrained while the
> aliphatic chains are not.  I then execute a 2000 ps simulation without
> restraints to allow equilibration followed by another 2000 ps simulation
> for
> the analysis.  Once this has been completed I take the resulting gromos
> file
> and use editconf to increase the box vectors so that a packing density of
> 19
> A^2/molecule is obtained.  Similar to the case of 18.5 A^2/molecule, a
> ps equilibration MD run is completed for the 19 A^2/molecule system
> followed
> by another 2000 ps simulation for analysis.  The resulting gromos file
> the second 2000 ps run for the 19 A^2/molecule system is then used via
> editconf to generate a box size yielding a packing density of 19.5
> A^2/molecule.  Again, a 2000 ps equilibration simulation is conducted
> followed by a 2000 ps analysis run.  This process is continued up to 25
> A^2/molecule.
> As you probably expect, I am trying to generate a Pi-A (surface pressure
> vs.
> area) isotherm by manipulating the area per molecule and monitoring the
> pressure.  The problem I have encountered is that the diagonal pressure
> values (Pres-XX, Pres-YY, and Pres-ZZ) fluctuate quite drastically from
> simulation at a given area per molecule to another.  For example when
> from 18.5 to 19 A^2/molecule, I see an increase rather a decrease in the
> diagonal pressure elements.  When going from 19 to 19.5 A^2/molecule I
> observe a decrease in pressure followed by an increase from 19.5 to 20
> A^2/molecule.  This occurs at all packing densities up to 25 A^2/molecule.
> The work described above, as noted, was performed in vacuum.  There has
> been
> some work cited in the literature in which an entire Pi-A diagram was
> obtained for a fatty acid monolayer.  Their system was in vacuum, however,
> they utilized a modified LJ potential in a plane just below the headgroups
> to simulate the interaction of the fatty acids with water (something my
> system is missing).  Alternatively, I have also looked at work by Alhstrom
> and Berendsen who simulated a lecithin monolayer (using a dual monolayer
> system) and monitored the surface pressure at a specified packing density
> which agreed quite well with experiments.  They employed a dual monolayer
> system with a layer of water in between the two monolayers (This was done
> to
> provide periodic boundary conditions in z).  I am in the process of
> conducting similar simulations, however, I first wanted to see if it is
> possible to find any success with a system in vacuum.
> Finally, to the question.  Does anyone have any tips or suggestions? 
> Below,
> I have provided some of the system parameters in the event that anyone is
> interested.  For example, I only generate velocities and have a
> start in the initial position restrained simulation.  The length of the
> in the z-direction is 7.5 nm (fatty acid length = 2.6 nm) so that the
> headgroups do not interact with the chains.  The methyl and methylene
> groups
> are represented by CH3 and CH2 groups in the GROMACS topology file.
> Periodic boundary conditions are imposed in xyz.
> Thank you,
> Roger McMullen

here are some tips:
-you could increase the simulation length. 
-if you are using Berendsen pressure bath, you could increase tau-p.
-you can simulate a bigger monolayer.
-what forcefield are you using?


Pedro Alexandre Lapido Loureiro
Laboratório de Física Biológica
Instituto de Biofísica
gmx-users mailing list
gmx-users at gromacs.org
Please don't post (un)subscribe requests to the list. Use the 
www interface or send it to gmx-users-request at gromacs.org.

More information about the gromacs.org_gmx-users mailing list