[gmx-users] umbrella sampling with distances larger than half box size

[anaome at fundp.ac.be]

Dear Users,

I am performing umbrella sampling simulations of the ring closure of  
DNA minicircles. The problem with pull_geometry=distance is that the  
end-to-end distance used as general coordinate becomes at some point  
larger than half the box size. The distance to the closest image is  
used in the histogram generation and WHAM procedure what is of course  
wrong. To avoid this problem I tried pull_geometry=direction_periodic  
with pull_vec1 taken as the vector between both ends of the DNA  
fragment and one end frozen but this set-up does not seem to work  
properly. Does anyone have a solution? Is there a possibility to  
implement a pull_geometry=distance_periodic in a next version of  
gromacs?

Thank you for helping!

Below are the relevant options used for umbrella sampling:
pull                     = umbrella
pull_geometry            = distance
pull_dim                 = Y Y Y
pull_start               = yes
pull_ngroups             = 1
pull_group0              = start
pull_group1              = end
pull_k1                  = 100


pull                     = umbrella
pull_geometry            = direction_periodic
pull_dim                 = Y Y Y
pull_start               = yes
pull_ngroups             = 1
pull_group0              = start
pull_group1              = end
pull_vec1=-0.4160005.3 -0.2650005.3 -2.4920005.3
pull_k1                  = 100
freezegrps               = start
freezedim                = Y Y Y



-- 
Aymeric Naômé
Ph. D. Student
UCPTS Division
University of Namur
61 Rue de Bruxelles
5000 Namur
BELGIUM