GROMACS version: 2025.2
GROMACS modification: No
Hi GROMACS community, I’m trying to construct a potential of mean force curve for two proteins in a lipid bilayer as a function of the distance between their centres of mass, using GROMACS with Martini forcefield. I’m trying to generate starting configurations for umbrella sampling by running a steered simulation, where the two proteins are initially at a distance of 10nm and I apply a pulling force to bring one of the proteins towards the other. I’m applying a gentle position restraint in the z direction on the pulled protein to keep it in the plane of the membrane.
However, I’m finding it hard to avoid membrane artefacts during the pull. If I take the pull really slowly, the region of membrane between the proteins buckles, and if I take it a bit quicker (even if I pause periodically to equilibrate), I end up getting a spike of membrane forming near where the proteins meet.
Any suggestions on how to set this up? Or recommendations from the literature? (I haven’t been able to find anything). Would be keen to avoid doing this as an all-atom simulation to keep the compute feasible.