[gmx-users] constraint pull code / gromacs 5.0.4

CARDELLINI ANNALISA d021974 at polito.it
Tue Jun 23 19:53:13 CEST 2015


Dear gromacs users,
I'm using the pull code to evaluate the Potential of Mean 
Forces between two nanoparticles(NPs) of Radius = 1nm.
Specifically, in order to create the configurations 
(i.e.the 2 NPs at different centr of mass distance) I used 
the constraint pull code.
Then for each configuration I run a simulation of 2ns. 
Part of the mdp file that I used is:
...
; 9) NEIGHBOR SEARCHING
; cut-off scheme
cutoff-scheme       = Verlet
nstlist             =  10    ;
ns_type             =  grid
pbc        =  xyz
periodic_molecules  =  yes
rlist               =  1.3     ;

; 10) ELECTROSTATICS
coulombtype         =  PME-Switch   ;
pme_order           =  4
rcoulomb            =  1.3     ;
rcoulomb-switch     =  1.25     ;

; 11) VdW
vdw-type            =  Cut-off  ;
rvdw-switch         =  1.3      ;
rvdw                =  1.3
DispCorr            =  EnerPres ;

; 13) EWALD
fourierspacing      =  0.12  ; (0.12) [nm] default
fourier_nx          =  0     ; (0) default
fourier_ny          =  0     ; (0) default
fourier_nz          =  0     ; (0) default
ewald_rtol          =  1e-5  ; (1e-5) default
optimize_fft        =  yes

; 14) TEMPERATURE COUPLING
Tcoupl                  = Nose-Hoover
tc_grps                 = UNK        SOL
tau_t                   = 0.1        0.1
ref_t                   = 300        300

; 15) PRESSURE COUPLING
Pcoupl              =  Parrinello-Rahman
Pcoupltype          =  isotropic
tau_p               =  2       2        ; (1) [ps] default
compressibility     =  4.5e-5  4.5e-5   ; [bar-1]
ref_p               =  1    1    ; [bar]
refcoord_scaling    =  com    ;

; 17) VELOCITY GENERATION
gen-vel                  = no
gen-temp                 = 300
gen-seed                 = -1

; 18) BONDS
constraints             = ; convert all bonds to 
constraints
constraint_algorithm    = ; LINear Constraint Solver
continuation            = yes           ;
lincs_order             = 4             ;
lincs_iter              = 1             ;
lincs_warnangle         = 90            ;

; Pull type:
pull                     = constraint                    ;
pull-geometry            = distance                      ;
pull-dim                 = Y N N                         ;
pull-start               = yes                           ;
pull-ngroups             = 2                             ;
pull-group1-name         = 1UNK                          ;
pull-group2-name         = 2UNK
pull-coord1-groups       = 1 2
pull-coord1-rate         = 0.0                           ; 
0.01 nm per ps
pull-nstxout             = 1000      ; every 2 ps
pull-nstfout             = 1000      ; every 2 ps
pull-coord1-init         = 0.0       ;

My problem is related to the resulting forces that I 
obtained. They do not seem to converge and they are really 
fluactuating for any configuration (i.e. indipendently 
from the com distance)
An example of pullf.xvg that I obtained at 2 nm com 
distance
..
1992.5000	-2367.46
1993.0000	-6738.59
1993.5000	361.472
1994.0000	647.644
1994.5000	-6222.61
1995.0000	6018.37
1995.5000	1921.95
1996.0000	-1618.72
1996.5000	1279.2
1997.0000	1302.08
1997.5000	2256.82
1998.0000	1485.05
1998.5000	2318.51
1999.0000	-3588.26
1999.5000	979.835
2000.0000	-678.607

Of course, the PMF that I got with a simple intagration of 
the forces is characterized by enormous error bars and its 
shape does not have any sense.

Do you have some suggestions for getting converged forces? 
Do you see some errors in the mdp file?
Thank you in advance for your help.
Annalisa



Annalisa Cardellini, M.Sc.
Ph.D. student

Politecnico di Torino - Energy Department
Corso Duca degli Abruzzi, 24
10129 Turin - ITALY
Tel: +39-011-090-4495
Fax: +39-011-090-4499

E-mail: annalisa.cardellini at polito.it
SMaLL: http://www.polito.it/small


More information about the gromacs.org_gmx-users mailing list