Hi. I am using Martini force-field and Gromacs package to simulate a protein-ligand complex (220 residues). The tutorial given for protein in Martini homepage doesn't say anything about charge neutralization. Since my system has a net -3 charge, I've neutralized it prior to minimization. But in the equilibration step, the job halts indicating some segmentation fault. Kindly suggest on how to interpret the error and trouble shoot the same. Thank you
First check the minimization. If the max force is high (more than 1000), you have to run again the minimization changing parameters and fixing your structure.
If the minimization is OK, you can start the equilibrium stage from a very low temperature (1K for example) and increase it slowly.
Most of our problems with " segmentation faults" are due to trying to run a system which is too big on a machine with too little memory. Give that a try ( a smaller system on a bigger machine and maybe with a smaller tilmestep). Sometimes the glaringly obvious is the solution.
Hi,
does it write pdbs for steps? If it does then system has become unstable, which suggests that either (1) your system was insufficiently minimized, (2) your .mdp settings are incorrect, and/or (3) the ligand topology is flawed. If it doesn't write pdbs then above answer might help.
When system tries to access the unallocated memory, it results in segmentation fault. So you should follow the suggestion by Muller.
You can try smaller time step for early stages of equilibrium, then increase it gradually.
Please take a look at you "min.log" file as well to make sure you do not have a high force point (always for MARTINI, force should be below 1000 or 500 to avoid blow up during equilibration stage). If the force is high, take a look at your initial conf (*.gro) file for the mentioned atom with high force to figure out what is wrong in that region of the structure.
Hope that helps.
First check the minimization. If the max force is high (more than 1000), you have to run again the minimization changing parameters and fixing your structure.
If the minimization is OK, you can start the equilibrium stage from a very low temperature (1K for example) and increase it slowly.
Hi
Neutralization of any system is an important step in simulation. The reason why they have skip the neutralization is because the system which they have taken is already neutralized, due to which they have skip that step.
And Segmentation Fault is due many reason like:
1. System is not minimized properly.
2. equilibration protocol (.mdp file) is not correct.
3. Memory/space problem,
Out of all better you can check the error and note when you are running any step.
Good Luck
Thank you all. I have another query. What should one do if maximum force is not attained? I tried increasing the number of steps to 100 from the default 10, for a system of 220 residues, but it doesnt help. Kindly suggest.Thanks in advance.
Hi, Is this occurring in energy minimization ?You can try to increase the number of steeps further or you can change the integrator i.e. minimization algorithm like cg, l-bfgs instead of steepest descent (steep).
As I mentioned previously,
1. reduce time step, e.g. from 200 fs to 50 or even less. Are you equilibrating in NVT or NPT? I suggest using NVT with very short time step, and increase it gradually. Then, switch to NPT.
2. If it did not help, check the atom with high force in VMD. You may need to change something in your initial conf.
3. Sometimes you put an inappropriate box size for your system. You can use editconf to change the box size to avoid having any atoms clashing with edges.
4. Look carefully your water molecules. A few of them might trapped or isolated from bulk water. This also causes explosion.
- Badges
- Science topic