[gmx-users] Implicit solvent simulation of protein folding

[Tobias Sikosek]

Hi everyone,

I’m trying to run protein folding simulations with implicit solvent and the Amber99sb-ILDN forcefield. 

My problem is that I cannot get a stable simulation after energy minimization. Either mdrun stops after lots of LINCS warnings or the simulation finishes but the atom coordinates are “infinite”. I have tried different parameter settings, some of which I found on gmx-users. My most recent .mdp files I have attached below.

I’ve used versions 4.5.5 and 4.6.3 of Gromacs.

The initial (unfolded) structure of ~60 residues was generated from an extended beta-strand followed by 1 ns of explicit solvent MD (after EM, NVT, NPT). 

I’m hoping it’s something obvious, since I have started using Gromacs fairly recently and so far have only used explicit solvent.  Any suggestions are welcome.

Cheers,
Tobias

-----------

em.mdp

integrator = steep
emtol = 1
emstep          = 0.01 
nsteps = 300000
constraint_algorithm = lincs 
constraints = none 
comm_mode = ANGULAR
tc-grps         = system       
tcoupl          = V-rescale
tau_t           = 0.1  
ref_t           = 350  
nstlist = 0 
ns_type = simple
rlist = 0
coulombtype = Cut-off
rcoulomb = 0 
rvdw = 0 
pbc = no 


md.mdp

integrator = md
nsteps = 5000000.0 
dt = 0.002
nstxout = 1000 
nstvout = 1000
nstxtcout = 1000
nstenergy = 1000
nstlog = 1000 
continuation = no 
constraint_algorithm = lincs
constraints = h-bonds 
lincs_iter = 1 
lincs_order = 4 
ns_type = simple 
nstlist = 0 
rlist = 0 
rcoulomb = 0 
rvdw = 0 
coulombtype = Cut-off
pme_order = 4
fourierspacing = 0.16
tcoupl = V-rescale
tc-grps = system
tau_t = 0.1 
ref_t = 350 
pcoupl = no
pbc = no
gen_vel = no 
comm_mode = ANGULAR
comm-grps           =  Protein
implicit_solvent = GBSA
gb_algorithm     = OBC 
nstgbradii       =  1.0
rgbradii = 0
gb_epsilon_solvent = 80
gb_dielectric_offset = 0.009
sa_algorithm         = Ace-approximation
sa_surface_tension   = 0.0054