[gmx-users] potential on non-protein molecules

[gianluca santarossa]

dear gmx,
    I just started to simulate non-protein molecules, and I'm dealing 
with some unexpected behaviors of my system.
By now, my system consists of an in vacuo simulation of mainly 
hydrophobic dendrimers - in this case poli-substituted benzenes bonded 
togheter by ammide bonds - in an NVT ensamble usign berendsen 
temperature coupling.
The main problem I have depends on huge fluctuactions of temperature 
(10% of the average value of 300K).

The topologies where created using prodrg, then I modified the 
parameters using gromacs protein forcefield. The structure consists only 
of benzenes and ammide groups, so I think parameters usually found in 
proteins should work. What do you think about that?

On the other side, can anyone suggest me an hydrophobic solvent I can 
use, so I can try an explicit solvent simulation?

Thanks
Gianluca