Hi everyone, I have been working on a protein-ligand MD simulation involving ligand molecules like cyclic sugars. During the equilibration I have used a separate temperature coupling group for the ligand. After running the simulation I have found that the mean ligand-temperature: 303.28 K is very close to the reference temperature: 303.15 K, but the instantaneous ligand-temperatures are fluctuating. The RMSD after calculating the temperature using gmx energy for ligand temperature is approximately 28 K. I have run the production MD for longer simulation time without any instability. Does this settings pose some kind of artifact into the properties or ligand binding abilities? Any help is much appreciated.
Hello,
It should not be necessary to have separate coupling groups at all, and even if you use them they should be of significant size so that the temperature of the particles can be properly distributed over the whole group. Just having a single ligand as a coupling group is asking for trouble.
Cheers
Paul
P.s.: I’m sure Berk can elaborate more on this topic if needed.
Thank you for your response. I understand now that over very few number of atoms the temperature coupling can become unstable. Does this affect the binding process of a ligand as well?
It will affect the temperature and thus velocities of the atoms of your
ligand, so I would expect there to be an effect, but I haven’t run such
calculations myself in a while.
Better to ask Justin or Berk :)
I am 100% with Paul here. Aside from extravagant cases aimed at for example probing heat transfer, it is hard for me to imagine the need for subsystem thermostatting. I say this while doing it for silly legacy reasons in a biological system, but every group here is like 200K atoms, so who cares…