How to determine if a system is well equlibrated

User discussions
1

Hi dear Gromacs developers!

I am using Gromacs 2021 + Charmm36 Jul21 to simulate a peptide amphiphile nanofiber system. I have tried various of combinations of parameters (number of peptides per layer, the distance between each layer, the concentration of salt, etc.) but the nanofiber breaks eventually. I am using semiisotropic and noticed that the box would expand by a large extend in production runs. The fiber seems it is being pulled by itself and would break in the middle. Since the box size is fluctuating a lot, I am assuming the system is not well equilibrated. I am doing the following to equlibrate the system:

2 ns NVT with position restraint fc = 1000
2 ns NPT with position restraint fc = 1000
2 ns NPT with position restraint fc = 600
2 ns NPT with position restraint fc = 300
2 ns NPT without position restraint.

I am using parameter similar to those in the membrane tutorial (KALP15 in DPPC).
How can I make sure the system is well equilibrated? The gromacs tutorial uses temperature.
Or is there any possibility causing the problem?

Best,
Yang

2

For membranes, it is potential energy and area per lipid. If it is stable it’s ok You don’t have a membrane, so you can’t compute area per lipid, but you can check potential energy I think. If the potential energy is stable and is on the same level (not increase or decrease too much your system is probably well equilibrated). I equilibrated my MGDG membrane (50 k atoms) after 100 ns and DGDG membrane (100k atoms) after 600 ns, but I don’t have any idea how much time you need for your system.

3

If you look at Justin’s Lysosomes in water and other gromacs tutorials, he has provided some basic idea of energy minimization and equlibration. https://mdtutorials.com/gmx/lysosome/06_equil.html

minimization:
the potential energy is expected to be negative;
second is check is maximum force, with a target value no more than 1000kj/mol.
note: thought one can arrive at a reasonable potential energy with F-max > 1000, that indicates an unstable simulation (gromacs doc 2018 : runtime errors)
EM ensure we have a reasonable starting structure, in terms of geometry and solvent orientation.

From what you say, it appears that minimization itself wasn’t ended with stability…

(Gromacs tutorial)
As for EQ: assumping you did NVT, NPT,
hen you can use temperature (expected to be plateau), pressure and density plots to check if the system is adequately EQed. First you can check if the system was brought to the desired temperature (which you set), then pressure contributes to bring the system to proper density (stabilized).

have you provided enough atoms, water to pack the box (ratios wise)?