[gmx-users] lipid simulations

[chris.neale at utoronto.ca]

> what parameters should I use for lipid simulations

That depends on the forcefield that you are using. I will tell you  
about what I do but this is a poor substitute for what you should  
really be doing, which is:

1) find the initial paper that publishes the forcefield that you are  
using and figure out how they used it
2) Find some very recent papers that use this same forcefield and  
figure out how they use it and then read the relevant papers that they  
reference in their methods section.

That said, I use the forcefield having a united atom representation of  
lipids originally proposed by Egberts et. al. (1994), with the charges  
from Chiu et. al. (1995), adjusted by Berger et. al. (1997), and with  
the additional scaling of 1-4 coulombics introduced by Lindahl and  
Edholm (2000). You can download these topologies from Peter Tieleman's  
website:  
http://moose.bio.ucalgary.ca/index.php?page=Structures_and_Topologies

It is common to use a twin range cutoff with LJ calculated every step  
up to 1.0nm and up to 1.4nm every 10 steps when the nonbonded list is  
updated. Coulombics are in real space below 1.0nm and by PME beyond.  
Pressure coupling is a semiisotropic berendsen barostat with tau_p of  
4.0ps. Temperature is coupled to a berendsen thermostat separately for  
i)lipids ii)water+ions iii)protein witha tau_t of 0.1ps. I use lincs  
to constrain all-bonds.

Note that I personally reduce the short range interaction cutoff to  
0.9nm from 1.0nm. But you would be better to stick with parameters  
that are in more common usage.

The protein forcefield that I use is opls-aa and I use tip4p water.  
When combining these forcefields, I use an unpublished method to scale  
1-4 coulombics differently for protein and lipid:
http://www.gromacs.org/pipermail/gmx-users/2006-September/023761.html
There are also a variety of related posts that come after that one but  
I can't remember what they are. You are going to need to do some  
searching.

> how to insert the protein in that bilayer.

You can either use the make_hole version of mdrun available for  
download from the user contributions section of the gromacs web site  
of a program called inflategtro available from Peter Tieleman's  
website: http://moose.bio.ucalgary.ca/index.php?page=Translate_lipdis