[gmx-users] POPC membrane simulations

[Chris Neale]

I didn't realize that DPPC was available from the G53a5 force field. Probably you should find out how that was parameterized first (You could start by looking at the paper that published G53a5 or find a paper when somebody expanded that ff to include DPPC). You would then have the information that you need to create your own POPC. I would not recommend that you do this via pdb2gmx. Instead, take the dppc.itp that you already have (Hopefully you have one that is included in a .top file instead of a .top file that is full of hundreds of copies what could be .itp parameters... if not then I would generate a 1 molecule dmpc.itp first, include it in a .top and run your dppc simulation to be sure that you have done it correctly).

While the modification of dppc.itp to popc.itp seems relatively simple, the double bond makes it more difficult. Take a look at how this in handled in the charm lipids and in tieleman's parameters. I have previously posted to this list about not understanding the way that Tieleman's parameters handle the double bond (specifically the pairs around the double bond in the middle of RB dihedrals)
http://www.gromacs.org/pipermail/gmx-users/2006-September/023630.html
and an excerpt from another post of mine:
"I have yet to figure out exactly why things are 
treated as they are in the [ pairs ] section for two non-RB double-bonded 
carbons in the middle of a long chain of RB carbons :: more specifically I am 
referring to lipid acyl chains with a double bond (for example pope.itp from Dr. 
Tieleman)."
Unfortunately the CHARMM ff probably won't help you here as I doubt that it uses the RB dihedrals. 

Once you have the .itp file you will need to add a few parameter definitions at the top (the ones that aren't covered in ffG53a5bon.itp and ffG53a5nb.itp. I don't know what they are but you should easily determine this. Don't take those values directly from Tieleman's parameters -- his are probably from an earlier version of gromos. Again, refer to the original papers and then use ffG53a5 as a source for the parameters.

I have created new lipids (based on the ones from Tieleman's site, but all I have personally done is shorten or lengthen chains or change a choline to an ethanolamine. Introducing a double bond is significantly more difficult and I can make no claim that what I have suggested here is complete. It's my best first thoughts on the topic. Probably the rest is up to you. However, this list would benefit if you would post your procedure upon completion so that others with similar questions in the future could be guided by your work.
Chris.