[gmx-users] mix berger ff with opls

[chris.neale at utoronto.ca]

Dear Bin:

Read the in-depth procedure here:  
http://www.pomeslab.com/files/lipidCombinationRules.pdf

Note that the 0.125 scaling factor is not anything that you need to do  
in addition to the files that you can download from Tieleman's site --  
it is already included there (you can confirm this for yourself by  
looking at the topologies downloaded form there). This is because the  
0.125 scaling (1/8 scaling) of 1-4 LJ is part of the original Berger  
lipids.

Note that the "Berger" lipids were in fact developed by many people  
and they have come to be named by only one of the contributors. The  
reference that I use for my 1/8 scaling of the LJ 1-4 is:

Lindahl, E., and O. Edholm. 2000. Mesoscopic undulations and thickness  
fluctuations in lipid bilayers from molecular dynamics simulations.  
Biophys. J. 79:426?433.

where they state in the methods:

"1,4 electrostatic inter-actions were reduced a factor of 2 and 1,4  
Lennard?Jones interactions a factor of 8."

I believe that this idea originated in a Jorgensen paper looking at  
bulk simulations of n-alkanes, but ashamedly I can't recall for sure  
at this moment. If you need it, I can look it up next week (let me  
know).

BTW, the half-epsilon double-pairlist method is finally published. The  
reference is Biophys J. 2010 Mar 3;98(5):784-792. "An Iris-Like  
Mechanism of Pore Dilation in the CorA Magnesium Transport System."  
Chakrabarti N, Neale C, Payandeh J, Pai EF, Pomès R.

In our paper, I attempt to give a more complete description of the  
development of this lipid forcefield by writing:

The parameters for DMPC lipids (10) were the united atom  
representation first proposed by Egberts et al. (18), with charges  
from Chiu et al. (19) adjusted by Berger et al. (20), and with the  
additional scaling of 1-4 coulombic terms introduced by Lindahl and  
Edholm (21).

My description of the HEDP method in the referenced paper is thus:

We mixed the OPLS-AA protein force field with the Berger lipid  
parameters consistently by applying the half-epsilon double-pairlist  
method introduced in this work. This method avoids reparameterization  
(22) and retains the original parameters with greater fidelity.  
Specifically, the epsilon-values of the 1-4 Lennard-Jones (LJ)  
parameters of the lipids are divided by two in the pair types section  
of the GROMACS input file, and the list of 1-4 interactions is  
duplicated in the topology file of each lipid. The regular OPLS  
combination rules are then applied. In this way, the lipidic LJ and  
Coulombic 1-4 parameters are both cut in half and then included twice  
to yield properly scaled 1-4 interactions for both lipids and OPLS  
protein.

Hope this helps,
please post again to the list if you have further question.

Chris.



-- original message --

Dear all:

I was trying to convert the lipid.itp downloaded from Tieleman's
website to the format compatible with OPLS. I basically followed the
procedure in this post, which is really helpful:
http://www.mail-archive.com/gmx-users@gromacs.org/msg03459.html

But there is one comment there I don't understand:
after changing c6/c12
to sigma/epsilon. (gives effective fudgeLJ of 0.125).
   Can someone please explain what this means? Why do I have this
effective fudgeLJ?
Thanks a lot,Bin