grompp_direct_MD_lambda-FEP_bulk.mdp

; GROMACS configuration for ...eGFP simulations
;  virtual-sites NegCleGFP well-energy-conserving setup
; NPT simulation (amber99sb version and Slipids)

integrator               = md
tinit                    = 0
dt                       = 0.005  ; 5fs
nsteps                   = 100000000 ; 500ns
init_step                = 0

; Lambda-dynamics stuff
free-energy              = yes
delta_lambda             = 0.0 ; lambda won't change during simulation
dhdl_derivatives         = no
nstdhdl                  = 0
; For Gromacs 4.6.+
fep-lambdas =   0.000  0.092  0.183  0.271  0.358  0.441  0.522  0.597  0.669  0.734  0.794  0.849  1.0
init-lambda-state        = 12


; ELECTRIC FIELD
; E-z                      = 1 +0.1 0

nstxout                  = 0
nstvout                  = 0
nstfout                  = 0
nstlog                   = 1000   ; 5ps
nstcalcenergy            = 100    ; 0.5ps, default value=100 ; replex must be multiple of this value!
nstenergy                = 1000   ; 5ps
nstxtcout                = 2000   ; 10ps
xtc-precision            = 1000
xtc-grps                 = Protein
energygrps               = Protein Water_and_ions 
nstlist                  = 25 ; .05ps, Gromacs manual says values>20 give the best performance (for GPUs)
ns-type                  = grid ; grid search for neighbour list faster for larger sys. than simple
;Centre of mass movement removal
comm-mode                = Linear
nstcomm                  = 100 ; 1.0ps, This should suffice
comm-grps                = Protein Water_and_ions 
;nstcheckpoint            = 2000   ;10ps

; OPTIONS FOR ELECTROSTATICS AND VDW
cutoff-scheme            = verlet      ; group is alternative
coulombtype              = PME			; Particle mesh Ewald, do not change
rcoulomb                 = 1.0			; Real-space cut-off
vdwtype                  = cut-off
vdw-modifier             = potential-shift-verlet   ; =shifted-vdw with Verlet_cutoff_scheme, =none with grid_cutoff_scheme
rvdw                     = 1.0			; VdW cut-off, orig paper rvdw=1.4, but this value is still fine (tested for both rigbonds an VSites, almost identical)
DispCorr                 = EnerPres		; Dispersion corrections to both the potential and pressure
table-extension          = 1.0
fourierspacing		       = 0.12			; PME grid

; Options no longer required with Verlet cut-off scheme
;rlist			          = 1.0			; Short-range neighbor list
;rlistlong		          = 1.6			; Long-range neighbor list 

; OPTIONS FOR WEAK COUPLING ALGORITHMS 
tcoupl                   = v-rescale   ; Thermostat, v-rescale or nose-hoover is also fine
tc-grps                  = Protein Water_and_ions		; Couple lipids and Water_and_ions seperatly
tau-t                    = 1.0 1.0		; Time constant for temperature coupling
ref-t                    = 310 310		; Desired temperature (K)
Pcoupl                   = parrinello-rahman	; Barostat
Pcoupltype               = isotropic	; Pressure in xy and z couple separately
ref-p                    = 1.013		; Desired pressure (bar)
tau-p                    = 10.0		; Time constant for pressure coupling
compressibility          = 4.5e-5	; Same as for water

; CONSTRAINTS
constraints              = all-bonds
constraint-algorithm     = Lincs
lincs-order              = 6    ; Increasing these does not make much change to the precision of virtual sites (J. Melcr)
lincs-iter               = 2    ; 5,2 conserves energy in NVE well enough (J. Melcr)

; Energy minimization opts
emtol                    = 10
emstep                   = 0.01
niter                    = 20
fcstep                   = 0
nstcgsteep               = 1000
nbfgscorr                = 10 ; number of correction steps for LBFGS