[gmx-users] umbrella pulling, pullx vs gmx distance

[Joe Wu]

Hello experts,

First, I want to mention that I am using Gromacs 2018.1.

I was wondering what the differences between pullx.xvg and gmx distance are.

"pullx.xvg" is the output of an umbrella sample mdrun.
"distance.xvg" is the output of the following command (I am comparing the
z-component distance):

gmx distance -s umbrella01.tpr -f umbrella01.trr -oxyz dist.xvg -select
'com of group "Water" plus com of group "Ion"'

The general idea is I am restraining an ion at various heights from the
center of a water slab. I have 28 umbrella windows, each 0.05nm apart from
distances 0 to 1.4nm. When I compare the "pullx.xvg" and "distance.xvg"
output of any one of these windows, and it seems that for the most part,
both z-distances are the same (with a deviation of +/-0.001 due to
sig-figs); however there are a few time steps where the pullx distance
deviates a lot from the distance.xvg distance.

I am attaching a link to the graph comparing "pullx.xvg" and "distance.xvg."
https://drive.google.com/file/d/1Xc3_bChpMckxXdUqH6xkYD0IO5BW4pW3/view?usp=sharing


The pull portion of my MDP file is:

; Pull code

pull                    = yes

pull_ngroups            = 2

pull_ncoords            = 1

pull_group1_name        = Water

pull_group2_name        = Ion

pull_coord1_type        = umbrella

pull_coord1_geometry    = direction

pull_coord1_vec         = 0 0 1

pull_coord1_groups      = 1 2

pull_coord1_dim         = N N Y

pull_coord1_rate        = 0.0

pull_coord1_k           = 4000

pull_coord1_start       = no

pull_coord1_init        = 0.1

Does anyone know why these two z-component distances are not identical, as
they should? I am graphing both "pullx.xvg" and "distance.xvg" z-distance
as a function of time; both graphs should be superimposable, but they are
not. Why is this the case?

Please let me know if you need more information.

Thank you for your time!

Best,
Joe