Hi everyone,
I’m running a simulation of a peptide in water and have a few questions about setting up the MDP parameters. Just to note, I’m using the V-rescale thermostat and C-rescale barostat.
From what I’ve seen on the forums, the recommended values for tau_t are 1 ps and tau_p are 5 ps.
As far as I understand, these values are for the production run. Are they suitable for a small system with a 30-amino-acid peptide? I’ve also come across studies where tau_p = 2 was used.
Also, if these values are for production, what should be the values for equilibration? Should they be smaller? For example, I’ve seen tau_t = 0.1 used during NVT equilibration.
Those values should be suitable for a small peptide as well, yes. Basing your values on other studies might not be very good (unless you are trying to reproduce them), since previous poor choices of parameters may remain in use for no good reason. You should know why you choose the settings you use.
In general, you can use a lower tau_t for equilibration, to have a stricter control of the temperature. You would usually be able to use a lower tau_p as well, but setting it too low may make the simulation unstable, even when using first order coupling algorithms.
I would think that tau_t between 0.5 and 2 ps and tau_p between 2 and 10 ps (with a c-rescale barostat) would work for most atomistic systems. But is there a specific reason why you think you need to set them lower than 1 and 5 ps, respectively?
I’ve often seen tau_p = 5 ps used for larger systems, such as membranes or proteins, which led me to hypothesize that it might not be optimal for smaller systems like peptide simulations. However, I’ve understood that first-order coupling algorithms are less sensitive to tau values, which makes this less of a concern.
Large systems usually have a more stable pressure, so I would think that smaller systems might need a higher tau_p, rather than lower, if it would need to be changed at all. But I might be wrong in that assumption.