[gmx-users] Cylinder pulling through bilayer

[Gmx QA]

Dear list

I am attempting to pull a small molecule though a bilayer using the pull
geometry cylinder with gromacs v 2016.

This is the relevant portion of my mdp-file:

pull                      = yes
pull-ngroups             = 2
pull-ncoords             = 1
pull-coord1-groups       = 1 2
pull-group1-name         = LIG
pull-group2-name         = MEM
pull-coord1-type         = umbrella
pull-coord1-geometry     = cylinder
pull-coord1-rate         = 0.1
pull-coord1-vec          = 0 0 1
pull-coord1-k            = 1000
pull-coord1-start        = yes
pull-coord1-init         = 0
pull-cylinder-r          = 1.5

The pull-rate is very fast because I'm only doing preliminary test. At the
start, the drug molecule is in -z position compared to the membrane.

When doing grompp:

$ gmx grompp -f umbrella_md_test.mdp -c npt.gro -p topol.top -o
pull_test.tpr

The output makes no sense:
Using a fourier grid of 72x72x192, spacing 0.113 0.113 0.111
Pull group  natoms  pbc atom  distance at start  reference at t=0
       1        36        18
       2     32500     64286        -nan nm           -nan nm
Estimate for the relative computational load of the PME mesh part: 0.44
This run will generate roughly 14 Mb of data


I.e. nan's for distance. If I however switch in the mdp file so that
pull-group1-name = MEM and pull-group2-name = LIG, the distance gets
correctly calculated. But this does not seem to be what is prescribed in
the manual for cylinder pulling, where is says that the cylinder is formed
from the first group (should be the drug molecule) and through the com of
the reference group (the membrane in my case),

I think there is something I am missing?

Thanks!
/PK