[gmx-users] Lipids running out of simulation box

[徐以会]

Dear all,

I am simulating a lipid bilayer under dehydration (1:5 lipid/water). In order to compare it with the bilayer periodicity d of X-ray diffraction. As the paper suggests for dehydrated membrane,
https://www.researchgate.net/publication/292341967_Solvent-exposed_lipid_tail_protrusions_depend_on_lipid_membrane_composition_and_curvature
I built up a double-bilayer using packmol rather than solvate (to precisely control the number of water).
As the water layer is very thin, many DOPC run out of the simulation box.  Trjconv can surely repair the appearance, but I assume the bilayer periodicity (z1-z2) is longer accurate. 

So my questions:
(1) Is double bilayer a good system to obtain d?
(2) Does lipids out of box affect the calculation of d?
(3) If so, how to prevent it?  

Kind regards,
Yihui Xu
---
School of Physical Science and Technology at ShanghaiTech University
Shanghai, China