[gmx-users] Umbrella sampling - regd

[ramesh cheerla]

Dear Gromacs users,

                                     I am performing  umbrella sampling
simulations by selecting configurations with spacing window of  0.02nm from
pulling simulations with the following options in .mdp file to get PMF
along a axis for permeation of ion through a channel.

; Pull code
pull            = umbrella
pull_geometry   = position  ; simple distance increase
pull_dim       = N Y N
pull_vec1        = 0 -1 0
pull_start      = yes       ; define initial COM distance > 0
pull_ngroups    = 1
pull_group0     = FIXEDATOM
pull_group1     = SMDATOM
pull_rate1      = 0.0     ; 0.01 nm per ps = 10 nm per ns
pull_k1         = 2000      ; kJ mol^-1 nm^-2
pull_init1      = 0

And   am using g_wham to get PMF, while using g_wham,  am getting the
warnings like :
WARNING, no data point in bin 139 (z=-0.492557) !
You may not get a reasonable profile. Check your histograms!
Warning, poor sampling bin 193 (z=-0.364322). Check your histograms!
Warning, poor sampling bin 194 (z=-0.361947). Check your histograms!
Warning, poor sampling bin 195 (z=-0.359572). Check your histograms!
Warning, poor sampling bin 196 (z=-0.357197). Check your histograms!
Warning, poor sampling bin 197 (z=-0.354823). Check your histograms!
Warning, poor sampling bin 198 (z=-0.352448). Check your histograms!
Warning, poor sampling bin 199 (z=-0.350073). Check your histograms!
Initialized rapid wham stuff (contrib tolerance 6.66667e-08)
Evaluating only 2115 of 3000 expressions.
I have checked my histograms and PMF profile, by doing so I have understood
that histograms overlapping is restricted to only some regions and there is
no histograms at all in some regions i.e severe sampling problems at
barrier regions.  here I am sending the link that containing PMF profile
that I have obtained,  http://researchweb.iiit.ac.in/~bipin.singh/pmf.jpg

Can anybody tell me how can I get proper sampling at barrier regions is
there any efficient way to do this, how about meta dynamics with suitable
CV (collective variable) ? .


Thank you in advance.


Regards,
Ramesh.