AWH reaction coordinate

GROMACS version: 2018
GROMACS modification: No

Hi all,
I have a protein with two reacting groups. The distance between them observed in the x-ray structure is considered too long for that reaction to occur. So, a movement of the domains has been proposed to shorten that distance down to a widely accepted threshold.
The distance between the two groups intuitively seems a good choice for a reaction coordinate to describe this hypothetical conformational change.
I’m studying the AWH method, and I would like to use it to estimate the free energy of the process but it doesn’t allow me (with my settings) to bring two groups closer.
When I feed grompp with the following mdp parameters

; PULL CODE
pull = yes
pull-ngroups = 2
pull-group1-name = group1
pull-group2-name = group2
pull-ncoords = 1
pull-coord1-type = external-potential
pull-coord1-geometry = distance
pull-coord1-groups = 1 2
pull-coord1-potential-provider = awh

; AWH CODE
awh = yes
awh-potential = convolved
awh-nbias = 1
awh1-ndim = 1
awh1-dim1-coord-index = 1
awh1-dim1-start = 1.85
awh1-dim1-end = 1.35
awh1-dim1-force-constant = 150000
awh1-dim1-diffusion = 5e-5
awh1-error-init = 5
awh-share-multisim = yes
awh1-share-group = 1
awh1-equilibrate-histogram = yes

it prints this:
Fatal error:
For the non-periodic pull coordinates awh1-dim1-start cannot be larger than
awh1-dim1-end

Any suggestions on how to treat the R.C. geometry?

Many thanks in advance,

Fabio

Hello,
I stumbled upon the same problem. I just can’t pull 2 groups towards one another in a non-periodic system, which makes use of awh impossible for some of my systems. Have you found a solution? Should we proceed with a bug report/feature request?
Regards,
g

Have you tried with pull-geometry = direction instead?