Re: [gmx-users] Smaller Area Per Lipid for DPPC Bilayer

[Thomas Piggot]

Hi,

While comparing to other simulations can be useful, I would argue that 
the real test for the combination of force field and simulation 
parameters is to determine if the simulated membrane properties compare 
well to the experimentally determined values. As long as they do this, 
you can argue that your choices are sensible. Obviously if you are 
including proteins (or other molecules) in the system, the parameters 
need to also be shown to work well for these too.

As for not including the dispersion correction, yes it is fine to do 
this (if it improves the behaviour of your membrane) as a dispersion 
correction is most appropriately applied to homogeneous systems and not 
membranes.

Regarding the large range of values seen, I would only be concerned if 
you are exactly reproducing what other people have done, in terms of 
force field and simulation parameters used and seeing large differences 
to what they report. As I mentioned before, fairly small changes in some 
of these parameters can make some pretty substantial impacts upon the 
membrane properties. You also should be careful to ensure that the 
properties of the membranes you are analysing are converged (a block 
analysis is the way to properly check this).

Cheers

Tom

On 12/09/12 18:21, David Ackerman wrote:
Hi,

Thank you for your response. As to my concern about incorrect areas
and diffusion, I am basing it off of other papers that simulate DPPC
bilayers.

For instance, in this paper:
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3251217/figure/F12/ , they
simulate a DPPC bilayer with DiI molecules in it. I did the same
simulation, but whereas they get APL of ~.64-.65 nm^2, mine are again
~0.03-0.04 nm^2 smaller. Also, in this paper they show that the lipids
diffuse ~1.1 nm^2 over the span of 20 ns, whereas I get a much slower
rate of traveling ~1.4-1.6 nm^2 over 90 ns. As mentioned in the other
response, if I turn off dispersion correction I get higher APL
(~.65-.66 nm^2) and diffusion values that more closely match this and
other papers.

These APLs and diffusion values are similar for some other papers that
simulate DPPC bilayers.

Is it ok to have ranges this large compared to these other
simulations, and does it make physical sense to turn off the
dispersion correction for this force field?

Thanks for your time,
David

On Wed, Sep 12, 2012 at 11:29 AM, Justin Lemkul <jalem...@vt.edu> wrote:

On 9/12/12 10:56 AM, David Ackerman wrote:
Hello,

I have been basing some DPPC bilayer simulations off of files from
Justin Lemkul's tutorial, including the .itp files and .mdp files.
Everything has been working fine except that my area/lipid seems to be
too low and my diffusion coefficient seems to be too slow compared to
experimental values. As a test, I just started with Tieleman's

How far off are the diffusion constants?  I have never had a lot of luck
reproducing experimental values, but this may reflect a limitation of the
parameter set, simulation length, or both.


equilibrated 128 DPPC bilayer system, including the waters, and ran a
simulation using the mdp file below (note though I selected
continuation=yes, this was in fact not continued from a previous
equilibration). The simulation has been running for ~75 ns so far, and
the area/lipid is on average ~.61-.62 nm^2 . When I do full

That sounds like the expected outcome for this force field.  Why do you say
that is too low?


temperature/pressure equilibrations, even using different
thermostats/barostats, I seem to get area/lipid values similar to
these. Also, my diffusion coefficients are smaller than those reported
in papers invovling DPPC bilayers. I was wondering what the possible
reasons for this could be. Any help you could provide would be great.

Curiosities in the .mdp file:


tcoupl          = Nose-Hoover   ; Less accurate thermostat
tc-grps         = DPPC SOL      ; three coupling groups - more accurate
tau_t           = 0.1   0.1     ; time constant, in ps
ref_t           = 323   323     ; reference temperature, one for each

Why is your tau_t so small?  Generally one should use 0.5 - 2.0 with
Nose-Hoover.


group, in K
; Pressure coupling is on
pcoupl          = Parrinello-Rahman     ; Pressure coupling on in NPT
pcoupltype      = semiisotropic         ; uniform scaling of x-y box
vectors, independent z
tau_p           = 1.0           ; time constant, in ps
ref_p           = 0.0 1.0               ; reference pressure, x-y, z (in
bar)

Why are you setting zero pressure along the x-y plane?


compressibility = 4.5e-5   4.5e-5       ; isothermal compressibility,
bar^-1
; Periodic boundary conditions
pbc             = xyz           ; 3-D PBC
; Dispersion correction
DispCorr        = EnerPres      ; account for cut-off vdW scheme
; Velocity generation
gen_vel         = no            ; Velocity generation is off

If you are not continuing from a previous run (as you say above) and you are
also not generating velocities, you may be delaying equilibration by
allowing the initial forces dictate the velocities.  I suppose if the run is
stable enough, this is not a huge problem, but in general this combination
is not recommended.

-Justin

--
========================================

Justin A. Lemkul, Ph.D.
Research Scientist
Department of Biochemistry
Virginia Tech
Blacksburg, VA
jalemkul[at]vt.edu | (540) 231-9080
http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin

========================================
--
gmx-users mailing list    gmx-users@gromacs.org
http://lists.gromacs.org/mailman/listinfo/gmx-users
* Please search the archive at
http://www.gromacs.org/Support/Mailing_Lists/Search before posting!
* Please don't post (un)subscribe requests to the list. Use the www
interface or send it to gmx-users-requ...@gromacs.org.
* Can't post? Read http://www.gromacs.org/Support/Mailing_Lists

--
Dr Thomas Piggot
University of Southampton, UK.

--
gmx-users mailing list    gmx-users@gromacs.org
http://lists.gromacs.org/mailman/listinfo/gmx-users
* Please search the archive at 
http://www.gromacs.org/Support/Mailing_Lists/Search before posting!
* Please don't post (un)subscribe requests to the list. Use the 
www interface or send it to gmx-users-requ...@gromacs.org.
* Can't post? Read http://www.gromacs.org/Support/Mailing_Lists