Hi all,
I have a couple of questions for PMF calculations in Gromacs. I placed a nanoparticle (NP) in the water region approximately 1nm away from the membrane and by using Gromacs pull code, I am pulling the nanoparticle in the z-direction through the bilayer center plane. However, the tutorial and articles imply that the reference group is treated with a position restraint in z-direction. Here is my pull code:
pull = yes
pull_ncoords = 1
pull-coord1-type = umbrella
pull-coord1-geometry = distance
pull-coord1-dim = N N Y
pull_constr_tol = 1e-06
pull-coord1-start = yes
pull_nstxout = 100
pull_nstfout = 100
pull_ngroups = 2
pull-group1-name = NP
pull-group2-name = MEMBRANE
pull-coord1-groups = 1 2
pull-coord1-rate = -0.01
pull-coord1-k = 1000
-I am wondering whether can I positionally restraint the whole membrane in z-direction? If so, how? Before, I have only done it for PO4 beads in the bilayer.
-While pulling, equilibration and umbrella simulation, should I separate the groups as NP Membrane and Water for center of mass motion removal (comm-grps)?
-In Justin's tutorial, COM pulling distance is in increasing order, but my reaction coordinate is between extracellular NP and membrane. Therefore center of mass distances decrease during pulling. I mean the COM distance does not go from 0 to 4nm but 4nm to 0 nm. Does it make any difference?
-For umbrella simulation, I took pull-coord1-rate = 0.0, however the NP not only moved in the window, in x and y direction but also in the z-direction. Can the little deviations of NP from the window be tolerated or do I need to increase the harmonic potential?
Thank you so much!
Hi,
You can apply positional restraints using the topology file (or .itp file). It’s explained here (http://www.gromacs.org/Documentation/How-tos/Position_Restraints). Applying restraints can be delusive since every restraint will create a bias or artifact in your simulation. At least these restraints should be minimized so they do not affect your system. You can check how PMF profiles are changing with these restraints since you are performing umbrella simulations. If your system is deformed without restraints, using collective variables may be another option. Additionally, you can apply a cylindrical potential to your NP to avoid the large fluctuations in the x-y plane. You can check the shape of this cylindrical potential here. (http://www.alioacar.rf.gd/scripts.html)
Cheers,
Ali
Thank you Ali Osman Acar but the documentation is about restraining an atom in a molecule. What I ask is whether I can restrain the whole molecule, the membrane, in z-direction or not. If possible, how it is written in Gromacs language? If not, maybe freezing the membrane can be an option.
You need to create a list of atom indices that belong to your molecule (or membrane) using 'gmx make_ndx'.Then 'gmx genrestr' can create a new restraint file with that list.
Or
Just edit the default restraint file (posre.itp) using a text editor. But if you have a lot of atoms, this is not the appropriate way to do this i guess.
Cheers,
Ali
Ali Osman Acar is right on two accounts. First, you can positionally restrain the entire membrane in the Z-direction by creating an index group for the membrane, generating a posre.itp file using gmx genrestr, including that posre.itp in your topol.top file, and defining that restraint in your .mdp file's "define = " line.
Second, he's also correct to caution you against applying a position restraint to your entire membrane. Doing so would likely alter the membrane binding dynamics of your nanoparticle.
Building on that, Gülşah Gül, I don't necessarily see why you you have to restrain the membrane, though I have seen some simulations where bilayers drift. Is your bilayer drifting during your steered MD run? In that case, maybe it is necessary because you'll want your window spacing to be consistent
If that's the case, and the nanoparticle only interacts with one leaflet, maybe you could try restraining the Z-position of lipid phosphates in the opposing leaflet. That may be sufficient to hold the membrane in place. You could do that using gmx select and selecting the name of your lipids, the name of your phosphate atoms, and saying that they're below a certain Z-height.
Alternatively, you could try placing a flat-bottom restraint on the opposing side of the bilayer to ensure your bilayer doesn't drift away as you pull the nanoparticle into the bilayer. You could also set this up to only act on lipid phosphates if you like. Any time the lipids drifted too far down in the Z-direction, they would basically hit a wall and be pushed back where they were. The radius and force constant could probably be fairly small in this example.
- Badges
- Science topic
- Similar topics
- Chemistry
- Theoretical Chemistry
- Computational Chemistry
- Molecular Dynamics