[gmx-users] on g_potential and lipid bilayer analysis

[Francesco Tordini]

Dear all,
i'm trying to simulate an asymmetric membrane, created by
editing by hand the dppc64 system found on Tieleman's site.
The first dynamic run was good, and now i'm calculating
some system properties. I was able to manage order parameters
(g_order) and atom densities (g_density), but i've noticed
a strange behavior of the gromacs g_potential tool. At first, i
addressed the problem to the complexity of my system.
To check this problem, i've simulated also the pure dppc64 sytem
in the same conditions of the mixed bilayer. Again the run was good,
but g_potential can't produce a symmetric profile for the potential
along the membrane z-direction. Again, the shape of the chart is very
sensitive to the number of slices used for the calculation.
For example, for the dppc bilayer, i obtain a symmetric chart
(as one can expect for a sytem like this) only with 151 slices. Little
variations of these numbers produce big changes especially on the values
of the potential in the two acqueous phases, desymmetrizing the data.
Maybe i'm missing something, but i've read for example (Tieleman &
Berendsen 1996) that the "data was symmetrized" for the potential chart.
Moreover, even if g_potential assigns 0 potential to the water phase,
there is a particular reason for the assignment of zero potential to the
middle of the bilayer (and than to scale the potential values on the y
axis, i suppose)? --> (again Tieleman & Berendsen 1996).
This aspect is important to me, especially for a correct interpretation
of the asymmetric bilayer properties.

Cheers

Francesco

------------------------------------------

Dr. Francesco Tordini
Università di Milano-Bicocca
Dipartimento di Biotecnologie e Bioscienze
P.zza della Scienza, 2 - 20126 Milano
Tel.  0264483473 / Fax. 0264483478
cell. 3282794451
email tordini at tiscalinet.it