[gmx-users] When to use Dispersion Correction for Lipid Bilayers

[David Ackerman]

Hello,

I was wondering when it is appropriate to use Dispersion Correction
for lipid bilayers, or which setting (no, EnerPres, or Ener) is best.
I have seen it used in some discussions, and not used in others. As
for myself, I have simulated a DPPC bilayer. With DispCorr=EnerPres, I
get an area per lipid of ~.615 nm^2 (whereas other reported values are
closer to .64 nm^2) and slower diffusion than is reported. However,
when I don't have a DispCorr term, my area per lipid becomes ~.656
nm^2 and my diffusion more closely matches other reported values.

So, should DispCorr be used at all for bilayer simulations, and if so,
which type is most appropriate?

Thank you for your help,
David