[gmx-users] TEE-REX problem

[Ramon Crehuet]

Hi gromacs users,

I am running TEE-REX calculations on a tetrameric protein with 3
replicas and I have come up several problems. I have calculated the
principal components of a 6 ns molecular dynamics. The temperature of
the essential subspace (T_es) in these replicas are 300, 400 and 650K
(tau_es = 0.05 ps), whereas the rest of the system is at 300K (coupling
constant of 0.1 ps).

1) T_es fluctuates in a huge range of temperatures (100-800K in the
replica at 400K and 200-1200K in the 3rd replica). I have read in the
T-Rex README file that the fluctuations of T_es are likely to be greater
than the temperature of the rest of the system but not greater than 40K,
so the problem is really serious. I cannot reduce these ranges by
reducing the coupling constant to 0.04 ps or by using a different
essential subspace.

2) This problem gets worse when I do the same calculation with a bigger
tetrameric protein. T_es gets aberrant values such as 20397713408 K.
Indeed, the T-Rex simulation of this protein required bigger tau_es
values to begin the simulation. If not, only the replica at 650 K
stopped at the first steps as a result of a huge separation between
adjacent atoms (system exploding warning). BTW, a traditional MD at 650K
has a normal behaviour.

Thanks in advance,

Ramon