[gmx-users] POPE lipid from Tieleman's web site, the box of xy dimension will shrink during the simulation

[Alex]

Though I doubt it will make a difference, it may be useful to put a space between the values in

> ref_p= 1.01.0        ; reference pressure, x-y, z (in bar)
> compressibility = 4.5e-54.5e-5; isothermal compressibility, bar^-1
> tau_t= 0.50.5;0.5        ; time constant, in ps

but in any case, this could be poor pre-equilibration prior to dynamic
simulations. Also, you have a relatively loose barostat in conjunction with
a loose thermostat (0.5 ps), so the initial shockwaves linger in the
system. Try the default constants of tau_p=2.0 and
tau_t = 0.1. Nose-Hoover thermostat (algorithmically oscillatory) does not help with long ref_t
either, maybe try v-rescale.

Alex


n> Dear all,
n>    I am new to lipid simulation. I  have downloaded the pope.pdb,
n> pope.itp, lipid.itp from Tieleman's web site
n> (http://wcm.ucalgary.ca/tieleman/downloads). And I have made a
n> directory named gromos53a6_lipid.ff followed Justin's
n> tutorial
n> (http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin/gmx-tutorials/membrane_protein/index.html).
n> Then I use pope.pdp to run a pure 340 POPE lipid MD at 310K. But I
n> find that the box of xy dimension will shrink from
n> 9.6nm*9.5nm to 9.0nm*8.9nm after 20 ns, then it fluctuates at that
n> value till 200ns. The  box of z dimension will increase a little to
n> hold the volume constant.  This will lead to the area per lipid decreased from 0.54 nm^2 to
n> 0.47nm^2.   At the same time, I find the order parameter of lipid
n> tail (Scd) is too big (above 0.3). Does any one know  where the problem is?

n>     When I use dppc128.pdb from Tieleman's web site  to run a pure
n> DPPC lipid MDat 323K, the box will not shrink. And the area per
n> lipid of DPPC is in accordance with literature, about 0.62nm^2. 


n>   The water model is SPC. My mdp file is also from Justin's
n> tutorial. Followed is my mdp file of POPE:


n> title= KALP15-DPPC Production MD 
n> ; Run parameters
n> integrator= md; leap-frog integrator
n> nsteps= 100000000; 2 * 500000 = 1000 ps (1 ns)
n> dt    = 0.002; 2 fs
n> ; Output control
n> nstxout= 0; save coordinates every 2 ps
n> nstvout= 0; save velocities every 2 ps
n> nstxtcout= 1000; xtc compressed trajectory output every 2 ps
n> nstenergy= 1000; save energies every 2 ps
n> nstlog= 1000; update log file every 2 ps
n> ; Bond parameters
n> continuation= yes    ; Restarting after NPT 
n> constraint_algorithm = lincs; holonomic constraints 
n> constraints= all-bonds        ; all bonds (even heavy atom-H bonds) constrained
n> lincs_iter= 1            ; accuracy of LINCS
n> lincs_order= 4            ; also related to accuracy
n> ; Neighborsearching
n> ns_type= grid; search neighboring grid cels
n> nstlist= 5    ; 10 fs
n> rlist= 1.2; short-range neighborlist cutoff (in nm)
n> rcoulomb= 1.2; short-range electrostatic cutoff (in nm)
n> rvdw= 1.2; short-range van der Waals cutoff (in nm)
n> ; Electrostatics
n> coulombtype= PME; Particle Mesh Ewald for long-range electrostatics
n> pme_order= 4    ; cubic interpolation
n> fourierspacing= 0.16; grid spacing for FFT
n> ; Temperature coupling is on
n> tcoupl= Nose-Hoover    ; More accurate thermostat
n> tc-grps= POPE SOL;SOL_CL; three coupling groups - more accurate
n> tau_t= 0.50.5;0.5        ; time constant, in ps
n> ref_t= 310 310;323        ; reference temperature, one for each group, in K
n> ; Pressure coupling is on
n> pcoupl= Parrinello-Rahman    ; Pressure coupling on in NPT
n> pcoupltype= semiisotropic    ; uniform scaling of x-y box vectors, independent z
n> tau_p= 2.0        ; time constant, in ps
n> ref_p= 1.01.0        ; reference pressure, x-y, z (in bar)
n> compressibility = 4.5e-54.5e-5; isothermal compressibility, bar^-1
n> ; Periodic boundary conditions
n> pbc    = xyz; 3-D PBC
n> ; Dispersion correction
n> DispCorr= EnerPres; account for cut-off vdW scheme
n> ; Velocity generation
n> gen_vel= no; Velocity generation is off
n> ; COM motion removal
n> ; These options remove motion of the protein/bilayer relative to the solvent/ions
n> nstcomm         = 1
n> comm-mode       = Linear
n> comm-grps       = POPE SOL 
n> ; Scale COM of reference coordinates (when postion restraint)
n> ;refcoord_scaling = com








n> And my topol file:
n> #include "gromos53a6_lipid.ff/forcefield.itp" 
n> #include "pope.itp" 


n>  ; Include water topology
n>  #include "gromos53a6_lipid.ff/spc.itp" 


n>  ; System specifications 
n> [ system ] 
n> 340-Lipid POPE and CNP
n>  [ molecules ] 
n> ; molecule name nr. 
n> POPE 340 
n> SOL   6729
n> Thanks in advance,

n> Nie Xuechuan
n>  Shanghai Institute of Applied Physics, Chinese Academy of Sciences



-- 
Best regards,
 Alex                            mailto:nedomacho at gmail.com