normal Martini dsRNA tutorial setup crashing for long runs.

  • ibrahim
  • ibrahim's Avatar Topic Author
  • Offline
  • Fresh Boarder
More
3 years 9 months ago - 3 years 8 months ago #8623 by ibrahim
Dear All,

I have been trying to simulate a relatively complex RNA structure using the CG martini forcefield. However, the simulation always crashed (even for just RNA in water) after few tens of nanoseconds.

I went back to the Martini ds-RNA tutorial, followed all the steps. Even for this simple case I see that the simulation is crashing after few tens of nanoseconds. I get LINCS warning, some bonds rotated more than 30 degrees and bond length becoming un-physically large exploding the simulation.

If I remove the stiff elastic network, it does not crash at least for the 100ns I simulated.
It seems there is a conflict between the elastic network potential and the angular potential defined in the itp file.

Does anyone faced the same issues. Could anyone suggest possible reasons. Is it a known issue for Martini RNA ?

I used the mdp files that are included in the tutorial

Thanks
Ibrahim
Last edit: 3 years 8 months ago by ibrahim. Reason: To add an important information

Please Log in or Create an account to join the conversation.

More
3 years 8 months ago #8628 by vainikka
I ran the ds-RNA tutorial just now, and didn't get any errors. Are you using the same .mdp files provided in the tutorial for your complex system?

If yes, then you might want to increase the steps in the energy minimization (it's only 100 steps in the tutorial) and increase the length of the equilibration run. Did you check whether your energy terms converge during min and eq steps?

Please Log in or Create an account to join the conversation.

  • ibrahim
  • ibrahim's Avatar Topic Author
  • Offline
  • Fresh Boarder
More
3 years 8 months ago #8629 by ibrahim
For how long did you ran it ? I see that with the equi.mdp, where the timestep is 10fs, it runs fine without crashing. But with 20fs timstep as in the mdrun.mdp, it crashes within 100ns.

Please Log in or Create an account to join the conversation.

  • ibrahim
  • ibrahim's Avatar Topic Author
  • Offline
  • Fresh Boarder
More
3 years 8 months ago #8630 by ibrahim
Here is how the Total energy and the LJ component of energy for RNA-RNA and Non-RNA-Non-RNA looks like during the equilibration for my tRNA system.
I seems converged.
drive.google.com/file/d/1_dtbs999F7NWJc9...iNT/view?usp=sharing

Also, Now for both the ds-RNA setup from the tutorial and my tRNA setup I did the following test (among many others):

1: Energy minimize with constraints replaced by stiff bonds.

2. Generate elastic network for the energy minimized structure with upper cutoff 1.0nm and lower cutoff 0.5 nm. This lower cutoff will probably alleviate any potential conflict between bonded/angular potentials and the elastic network potential.

3. If i proceed with the above procedure, I see that ds-RNA setup from the tutorial runs for about 540ns with timestep 20fs, before crashing.

5. With the same procedure my tRNA setup ran for about 340ns before crashing.

6. Also i set "Constraints = all-bonds" in the mdp.

Any further suggestion will be highly appreciated.

Thanks.

Here is the mdp parameter i am using:

title = Martini

integrator = md
dt = 0.02
nsteps = 50000000
nstcomm = 100
comm-grps =


nstxout = 0
nstvout = 0
nstfout = 0
nstlog = 10000;
nstenergy = 1000
nstxtcout = 10000
xtc_precision = 10000
xtc-grps =


energygrps = RNA Non-RNA;
cutoff-scheme = Verlet
nstlist = 20
ns_type = grid
pbc = xyz
verlet-buffer-tolerance = 0.005
refcoord-scaling = com
coulombtype = reaction-field
rcoulomb = 1.1
epsilon_r = 15
epsilon_rf = 0
vdw_type = cutoff
rvdw = 1.1

tcoupl = v-rescale
tc-grps = RNA Non-RNA ;
tau_t = 1.0 1.0 ;
ref_t = 323 323 ;

Pcoupl = berendsen ;
Pcoupltype = isotropic ;
tau_p = 8.0 ; 12.0
compressibility = 3e-4 3e-4 ;
ref_p = 1.0 1.0
gen_vel = no

constraints = all-bonds
constraint_algorithm = Lincs
continuation = no
lincs_order = 4
lincs_warnangle = 30

Please Log in or Create an account to join the conversation.

Time to create page: 0.094 seconds