GROMACS version:
GROMACS modification: Yes/No
Here post your question
Dear Comunity
I have a task to increase the number of lipids per area (or increase the density of the lipids). I have already created the system and relaxed it using CHARMM-GUI. My first thought was to take the relaxed system, apply pressure in a lateral direction(XX and YY), and set the pressure in Z direction 0. For that reason, I used the following code.
; Pressure coupling:
Pcoupl = c-rescale
Pcoupltype = semiisotropic
tau_p = 12.0
compressibility = 3e-4 0.0
ref_p = 1000.0 1.0
However, the pressure in all directions is increased (see attached pressure.png
)
So my question is how to increase pressure only in XX and YY directions, and most importantly, will that increase the density of the lipids, or do I need a more sophisticated method?
Thanks,
base_editor.
What do you mean by “I have a task”? What do you need to preserve, and does it have to be physical?
For many lipids, an easy way to increase the number of lipids per area is to decrease the temperature and let them crystallize/jellify. But I’m not sure if that counts as a solution for you.
By the way, the reason your setup didn’t work is most likely that you set compressibility in Z to 0, which has the effect of freezing the Z-dimension. But if you do it right, the lipids will not remain organized in a bilayer, don’t know if that counts either.
I am mimicking experimental results; there are two setups of lipids; in one case, the density of lipids is higher than in the second one. I want to simulate a system with two different densities at the same temperature.
It might be easier to build your systems in appropriate sized boxes for the required area per lipid values using something like packmol and run NVT simulations to keep the packing you want
Or at least with the x and y box lengths fixed so as to allow the amount of water you include to equilibrate properly in terms of the z box dimension size between membrane periodic images and don’t end up with vacuum bubbles in the box
I agree with @milosz.wieczor about the pressure increase in all directions. I think that you are decoupling the z length of the box by setting its compressibility to zero, you can check this by plotting the box side lengths from the .edr file. If this is the case then you are applying a positive pressure on x and y while keeping z fixed, and as such you pressure is exploding on z as well.
You are using CHARMM-GUI, so I guess you are using CHARMM36m forcefield? Why the compressibility of 3e-4? Also, 1000bar is a HUGE pressure, we are talking about roughly 1k atm, is there a reason why you picked such a value?
I am not an expert in these simulations, but since you want to change the density by changing the pressure at constant temperature, a natural setting seems to be (to me at least) a pressure-area isotherm study. There are a few examples here and there in literature (like this or this), and for sure the monolayer/liquid crystal field in physics is rich of such examples and has a ton of literature on everything (sims/theory/exp). It is worth taking a look to see if someone already did something similar so that you can copy the procedure or, if you are lucky enough and the systems they study are the same as yours, take a look at their isotherms and pick the values that already reproduce the densities/area per lipid that you want to have.
1 Like
If you’re trying to mimic experimental results, how did they achieve different densities? The most natural thing to do is to use two different lipids, AFAIK phospholipids containing MUFA/PUFA tails take up more area than ones with fully saturated tails.
Or maybe they manipulated the membrane using something like the Langmuir-Blodgett film technique, in which case you can mimic that by fixing the XY dimension at values larger than experimental.
The bottom line is, if you aim to reproduce experimental results, start by telling us how the experiment was realized. There are a thousand ways to do what you initially asked for, among which maybe a dozen will achieve what you indeed want to do.
Dear all
Thanks a lot for your replies.
So, regarding pressure. @milosz.wieczor was right about compressibility; I fixed that and was able to apply pressure. Nevertheless, that did not change the densities. @obZehn it is a martini system, that’s why 3e-4, and yes, 1000 was too big; it destroyed the lipid system. I decreased it incrementally to see the change. It either does not affect or deforms the surface.
Next, I will try to fix the x-y direction, add lipids, or do the simulations in the NVT ensemble (@TomPiggot suggestion).
I already asked my experimental counterparts to give me more details about the specifics, and I will let you know when I know.
Best,
base_editor.
With a big size combined with using a CG model then even fixing the box dimensions of the membrane may not work. This is because the membrane can easily curve/deform or you may even get budding. This may all depend on how good the force field is (or indeed other MD limitations), even if the experimental data says the lipid should be able to be compressed to this value.
As the others have said you really need to work out exactly what you are trying to do/test and why before deciding on how to best accomplish that.
