DispCorr parameter in GROMACS for Self Assembled Systems

GROMACS version: 2019.2
GROMACS modification: No

Dear GROMACS Users,

I had a question regarding the parameter, DispCorr. I am studying the self-assembly of surfactants in water. We anticipate that the surfactants will self-assemble into a mix of micelles and long rod-like aggregates. I am using the CHARMM36 forcefield to describe the surfactant molecules. In the manual, I came across this sentence, “Note that dispersion correction should be applied in the case of lipid monolayers, but not bilayers”.

I also observed that in the nvt.mdp file given for the Protein-Ligand tutorial on (http://www.mdtutorials.com/gmx/complex/Files/nvt.mdp), the comment on DispCorr mentions that no correction is used for a proteins with C36. Since I am studying self-assembly, I initiate my system with randomly distributed surfactant molecules. So, my system does not begin with monolayers or bilayers, nor does it contain proteins. I am confused about what I should be setting DispCorr to.

Could you please guide me on whether or not I should be using dispersion correction for my particular system of surfactants in water?

Thank you,


Tej S. Choksi

Assistant Professor

School of Chemical and Biomedical Engineering

Nanyang Technological University, Singapore

Block N1.2, B1-18

62 Nanyang Drive

Singapore 637459
Office: +65 6316 8940

The only system type for which one would use dispersion correction with CHARMM36 is a phospholipid monolayer. It should not be used for anything else. The comment in the documentation specifically addresses bilayers vs. monolayers because it is a point of confusion among users, but should perhaps be better phrased that only monolayers require it. Nothing else does.


Dr. Lemkul,

I have a question regarding this. What if someone uses Dispersion correction incorrectly for CHARMM, what would be consequences as correction term is a constant offset?


I think it will depend on the system. For a membrane, in which LJ interactions dominate, you’d probably get systematically too-favorable interactions among the alkyl groups. For a protein or nucleic acid in water, the differences are probably negligible, but I don’t know if anyone has ever looked at that before.