[gmx-users] PMF using umbrella sampling

[Bob Johnson]

Hello everyone,
I'm trying to compute the PMF for pulling two proteins apart. I'm
planning on setting up several runs where the COM separation of the
proteins is constrained and the constraint force is measured over the
course of a long simulation. However, I'm having trouble doing this
because I'm unsure of the inputs in the .ppa file. I'm trying to do
this with Gromacs 3.3.3. Here is the .ppa file I'm using:

runtype          = umbrella
ngroups          = 1
reference_group  = knob
reftype          = com
pulldim          = Y Y Y
group_1          = car
k1               = 1000
pos1             = 0.0 0.0 28.8989219666

Here, there are two proteins: car and knob. The COM of car is located
at (0.0 0.0 28.8989219666) relative to knob. The problem I'm
experiencing is that the simulation results in the two proteins being
pulled apart instead of constrained at distance pos1. I believe that
I'm misinterpreting the inputs. Does anyone know the solution?
Thanks,
Bob