Re: Membrane Simulations

From: bo liu (
Date: Tue Oct 14 2008 - 22:09:09 CDT

2008/10/15 Anshu Bhatia <>

> Hi Namd users,

Hi Anshu,

> I am trying to simulate
> (a)POPE membrane and
> (b) a gramicidin channel in POPE,
> similar to the one described in protocol. However, I need clarifications on
> a number of points.
> Here is what I am doing stepwise and my doubts in every step.
> (1) Get a 100 by 100 patch from vmd membrane plugin.
> (2) Solvate it more using the solvate plugin and remove waters which are
> inside the lipid bilayer.

You may also want to neutralize the system and keep the ion concentration at
physiological level by adding ions. PME requrires the simulation box to be
charge neutral. Although in some cases NAMD doesn't complain an
unneutralized system, this does influence the efficiency of PME calculations
(you can try).


> (3) Minimize for some steps
> (4) Hold everything constant but for waters and run a NVT

Maybe you want to do NpT simulations first to get a reasonable configuration
for the production run.
After the volume of the simulations box becomes stable, you can change it to
NVT ensemble, if you want.

> The purpose of these 4 steps is to get the additional water from solvate to
> its right density. But I see the water
> height rising towards the middle which does not seem right. What could be
> the reason for this, and how should

What do you mean by "water height"? Anyway, you may want to add this to your
.conf file:

 COMmotion no ;# allow initial center of mass motion?
 zeroMomentum yes ;# remove center of mass drift due to PME?

> I modify my protocol to get a good equilibrated membrane?
> After reading a few threads in the archive I found out that people are
> melting lipid tails to get lipids to their experimental
> density and then run a constant area simulation.

Constant area simulation only makes sense when you know the area/lipid value
(from experiments) and you want to stick to it, otherwise, i think we'd
better rely on simulations... For your system, you may know a correct
area/lipid value for pure POPE, however, can you guarantee a correct value
for the POPE-Protein complexes?

> (5)For equilibrating only the membrane/water system itself should melting
> tails precede the MD simulation of whole system.
> If so in what ensemble should the melting done.

For equilibration run, NpT ensemble should be the correct choice for
membrane simulations.

> I think the assumption here is that the whole membrane behaves uniformly
> after it has been brought to its ideal density values.
> Again my question is the following:
> (6) If I want to see how far in space the protein affects the membrane,
> should I still be using a constant area patch? How
> should I proceed if I want to observe the variances in lipid properties
> across the span of the membrane.

I think you can study the interplay between the protein and the membrane by
running NPAT ensembles, if you want to abide by the experimental area/lipid
value. Someone argues that it is the "size-effect" that makes using NPAT
ensemble necessary for membrane simulations, however, It is possible that
although you can generate the same area/lipid values as compared to
experimental studies, you may mislead some other properties of the membrane.
Well, I am not sure whether NPAT simulation really merits, i guess someone
else could give us a clue?

> Thanks,
> Anshu.


Bo Liu

> --
> Anshu Bhatia
> Graduate Student,
> Tri I Computational Biology Program ,
> Weill Cornell Graduate School,
> NY, NY

-Liu bo
Computational Biochemistry&Biophysics, Nano-bio systems: MD method;
College of Chemistry and Chemical Engineering, Graduate University of
Chinese Academy of Sciences, Beijing P.R. China
Office: (86)-010-88233187
Home: (86)-010-88259765
Cell: 13426057875

This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:49:58 CST