[gmx-users] Umbrella Sampling Error in Gromacs 5.0

[Jianhui Tian]

>
>
> Date: Wed, 22 Apr 2015 19:23:32 -0400
> From: Justin Lemkul <jalemkul at vt.edu>
> To: gmx-users at gromacs.org
> Subject: Re: [gmx-users] Umbrella Sampling Error in Gromacs 5.0
> Message-ID: <55382D74.5070902 at vt.edu>
> Content-Type: text/plain; charset=windows-1252; format=flowed
>
>
>
> On 4/22/15 4:59 PM, Jianhui Tian wrote:
> >>
> >> Date: Wed, 22 Apr 2015 16:05:18 -0400
> >> From: Justin Lemkul <jalemkul at vt.edu>
> >> To: gmx-users at gromacs.org
> >> Subject: Re: [gmx-users] Umbrella Sampling Error in Gromacs 5.0
> >> Message-ID: <5537FEFE.6020300 at vt.edu>
> >> Content-Type: text/plain; charset=windows-1252; format=flowed
> >>
> >>
> >>
> >> On 4/22/15 3:54 PM, Jianhui Tian wrote:
> >>> Dear Users,
> >>>
> >>> I am trying to do a umbrella pulling of one lipid out of a bilayer
> using
> >>> umbrella sampling code. In Gromacs 5.0, there is a big change in
> umbrella
> >>> sampling implementation. I set my pull parameters like this:
> >>>
> >>> pull             = umbrella
> >>> pull_geometry   = cylinder  ; Pull in the direction of pull-vec.
> >>> pull_coord1_vec = 0.0 0.0 -1.0
> >>> pull_r1 = 1.0
> >>> pull_r0 = 1.5
> >>> pull_start = yes
> >>> pull_ngroups = 2
> >>> pull_ncoords = 1
> >>> pull_coord1_groups     = 1 2
> >>> pull_group1_name     = PullLipid  ; one lipid that I want to pull out
> >>> pull_group2_name        = RefLipids   ; lipids on the other leaflet of
> >>> bilayer
> >>> pull_coord1_rate = 0.001 ; 0.001 nm per ps = 1 nm per ns
> >>> pull_coord1_k         = 1000      ; kJ mol^-1 nm^-2
> >>>
> >>> When I have a small system, it can work but had and following error and
> >>> stopped:
> >>>
> >>> Fatal error:
> >>> Distance between pull groups 1 and 2 (2.707004 nm) is larger than 0.49
> >>> times the box size (2.758373).
> >>> You might want to consider using "pull-geometry = direction-periodic"
> >>> instead.
> >>>
> >>> I guess this is due to a small box size (about 5.52 nm in x and y).
> >>>
> >>> Then I double the system size and redo the umbrella sampling. I got the
> >>> following message with grompp:
> >>> ########################################################
> >>> Pull group  natoms  pbc atom  distance at start     reference at t=0
> >>>          1       130     23335
> >>>          2      3136     38129    -nan    -nan
> >>> ########################################################
> >>
> >> Something is messed up in the input coordinates.  If the reference
> >> distance is
> >> "not a number," then you need to diagnose what it is you've done to the
> >> coordinates before doing anything else.
> >>
> >> -Justin
> >>
> >
> > Thanks, Justin.
> >
> > But, what do you mean by coordinates messed up?
>
> No idea, but you should never see "nan" if everything is sensible.  There
> are a
> bunch of possibilities, but the ones I can think of would cause a fatal
> error.
>
> > I have a gro file for generating the tpr for umbrella pulling.
> > The same GRO file can be used to generate a tpr file for normal MD run
> > without any problems.
> >
>
> OK, so it's specific to something being interpreted by the pull code.  Try
> varying some settings to see if it illuminates anything.  Otherwise, you
> need to
> upload all of your files somewhere (as well as a list of all your commands,
> exactly as you gave them) for us to look into it.
>
> -Justin
>
> --
> ==================================================
>
> Justin A. Lemkul, Ph.D.
> Ruth L. Kirschstein NRSA Postdoctoral Fellow
>
> Department of Pharmaceutical Sciences
> School of Pharmacy
> Health Sciences Facility II, Room 629
> University of Maryland, Baltimore
> 20 Penn St.
> Baltimore, MD 21201
>
> jalemkul at outerbanks.umaryland.edu | (410) 706-7441
> http://mackerell.umaryland.edu/~jalemkul
>
> ==================================================


Hi Justin,

My complete mdp file is as follows:

include =
define =
integrator              = md
tinit = 0.0
dt                      = 0.002
nsteps                  = 2500000
nstlog                  = 5000
nstxout                 = 5000
nstvout                 = 5000
nstxout-compressed      = 5000
nstenergy               = 5000
energygrps = Lipids Water
;
cutoff-scheme           = Verlet
nstlist                 = 10
ns_type = grid
pbc = xyz
rlist                   = 1.2
coulombtype             = pme
rcoulomb                = 1.2
fourierspacing = 0.12
vdwtype                 = Cut-off
vdw-modifier            = Force-switch
rvdw_switch             = 0.8
rvdw                    = 1.2
;
tcoupl                  = v-rescale ;Nose-Hoover
tc_grps                 = Lipids Water
tau_t                   = 1.0     1.0
ref_t                   = 323.15 323.15
ld-seed                 = 1454538035
;
pcoupl                  = berendsen ;Parrinello-Rahman
pcoupltype              = semiisotropic
tau_p                   = 5.0     5.0
compressibility         = 4.5e-5  4.5e-5
ref_p                   = 1.0     1.0
;
constraints             = h-bonds
constraint_algorithm    = LINCS
continuation            = no
lincs_order = 4
;
nstcomm                 = 100
comm_mode               = linear
comm_grps               = System
;
refcoord_scaling        = com
;
gen_vel = no
gen_temp = 323.15

;;;;;;;;;;;;;;;;;;;;;
;;;;;;;;;;;;;;;;;;;;;
;;;   Pull Code   ;;;
;;;   Steered MD  ;;;
;;;;;;;;;;;;;;;;;;;;;
;;;;;;;;;;;;;;;;;;;;;
pull             = umbrella
pull_geometry   = cylinder  ; Pull in the direction of pull-vec.
pull_coord1_vec = 0.0 0.0 -1.0
pull_r1 = 1.0
pull_r0 = 1.5
pull_start = yes
pull_ngroups = 2
pull_ncoords = 1
pull_coord1_groups     = 1 2
pull_group1_name     = PullLipid
pull_group2_name        = RefLipids
pull_coord1_rate = 0.001 ; 0.001 nm per ps = 1 nm per ns
pull_coord1_k         = 1000      ; kJ mol^-1 nm^-2


I used a grompp command to generate tpr file for umbrella pulling:

grompp -f pull.mdp -c GRO -p TOP -n NDX -o pull.tpr

Then, I used mdrun to do the umbrella pulling:
mdrun -s pull.tpr ...

I do not know whether this will be enough for pinning the problem.
Otherwise, I can upload the files if there is a place for doing so.

Best,
J.