Linc and Nan error in CG simulation md production step

[Rudhra]

GROMACS version:
GROMACS modification: Yes/No
Here post your question

Hi, Sir

I am getting linc and nan error during my MD production step. If i run MD below 30 femtosecond, I couldn’t observe the much changes in protein ( only four frame was in vmd view).
Kindly help me to solve this error.
Thank you in advance

dynamic.mdp (4.5 KB)
equilibration.mdp (4.6 KB)
minimization.mdp (4.6 KB)
serotonin_cg simulation.txt (1.8 KB)

[MagnusL]

I would suspect it’s the Parrinello-Rahman barostat that is the problem, in turn indicating that the system was perhaps not fully equilibrated. You are only equilibrating for 10ps. If you need a 0.001 ps timestep during equilibration, I think you can make two equilibration stages. Run a second equilibration with a 0.010 ps timestep, starting from the end of the first one (do not generate new velocities), and run it for 1-5 ns.

I would also recommend using:

tcoupl                   = v-rescale
tc-grps                  = system 
tau_t                    = 1.0   
ref_t                    = 300 
pcoupl                   = c-rescale
Pcoupltype               = isotropic
tau_p                    = 5.0
compressibility          = 3e-4
ref_p                    = 1.0

in both the equilibration and production stages.

[Rudhra]

Thank you so much for your prompt reply, I will run md as per your suggestion, let you know the outcome soon sir.

[MagnusL]

Then I’d recommend trying running the following stages instead:
EM
NVT Equilibration, dt = 0.005, nsteps = 400000, position restraints on
NPT Equilibration, dt = 0.005, nsteps = 400000, position restraints on
NPT Equilibration, dt = 0.010, nsteps = 400000
NPT Equilibration, dt = 0.020, nsteps = 500000

[Rudhra]

Thanks for the reply, I will try and let you know, sir

[MagnusL]

If the box size in NVT is very far off from the box size at the pressure of interest, you might need position restraints in the NPT equilibration as well. And then one or additional NPT equilibration without restraints.

[MagnusL]

Now I saw from your mdp files that you had position restraints in the NPT equilibration as well. Then I cannot think of any reason why it would crash.

[Rudhra]

Hi sir

I thought to start my npt equilibration (5ps) with position restraint on, then proceed with off.
I will run npt equilibration without position restraint and with lower dt =0.001, and update you the outcome soon.

Thank you.

[MagnusL]

By the way, when running gmx grompp, are you using the same input coordinates (-c) as reference coordinates for the restraints (-r)? Otherwise, I guess the restraints could result in very large forces.

[Rudhra]

Hi, sir
No, I have attached my commandline txt here.
Should I have to start my nvt equilibration from scratch?
Thanks for pointing out my mistake.

[MagnusL]

I don’t think there’s any need to rerun previous simulations. Often it’s a good idea to use an experimental structure as reference. But if it deviates too much from the starting configuration in the simulation, it can be problematic. If you apply restraints and the simulation stage before was unrestrained (which seems to have been the case here according to the two mdp files), you risk large forces if you restrain to completely different coordinates.

I’d say you can try the NPT without restraints or using the last NVT equilibration as reference coordinates.

[Rudhra]

Hi sir

NPT equilibration step (500 ps) running smoothly without any linc error. I used my nvt equilibration (10 ns) coordinates for restraints.
Once simulation get over, I will update you.
Thanks for your prompt response.

[Rudhra]

Hello, sir
The production MD is currently running at 100 ns. Initially, I encountered a pressure scaling inaccuracy of greater than 1%. After I raised the tau_p to 7, everything is running error-free and smoothly.
Is it okay if I run production MD without any restraint? I also attached my mdp file for your perusal
Thank you
dynamic.mdp (1.0 KB)

[MagnusL]

Hi,
Good to hear it’s progressing well.

You can definitely run the production part without restraints, that’s what’s commonly done if there are no specific reasons to restrain anything.

But I would recommend first running an extra equilibration stage (perhaps 5 to 10 ns) without restraints, before running the production simulation. During that equilibration I’d also recommend testing if the system is stable with a 20 fs timestep. If so, you can use that in the production simulation as well.

[Rudhra]

Hi, sir

I will run a simulation as per your comments, and update you the outcome soon.

[Rudhra]

Hi, Sir

My simulation got over well. Right now , I have planned to build membrane again with extra lipid molecule and cholesterol.
If I run simulation only with popc and pope lipids, I am not getting coordinate mismatch error. But if I add cholesterol and dpg3 lipid and ran insane step, I am getting coordinate mismatch error (4208 atoms difference) between my topology file and coordinate file (system.fix.gro)

How to fix this error? is there any issues in my cholesterol and dpg3 itp files.

DPG3.itp.txt (2.7 KB)

martini_v3.0_sterols_v1.0.itp.txt (2.6 KB)

[MagnusL]

I would recommend to count the number of each molecule type in the coordinate file and update the topology file accordingly. I’d also recommended to write a script to help you count.

[Rudhra]

Ok, Sir, Thanks for the reply.