[gmx-users] g_cluster cutoff

[Kukol, Andreas]

Hi,

I am using g_cluster to analyse a trajectory (protein backbone):

g_cluster -f traj.xtc -s topol.tpr -g cluster.log -skip 10 -cutoff 0.5 -fit -method linkage

As a result I get almost hundred clusters. The plot of the RMSD against time for the same trajectory generated with g_rms shows a maximum RMSD just below 0.5 nm. Therefore, with a -cutoff of 0.5 nm, I would expect to find that the whole trajectory clusters into ONE cluster. Am I expecting something wrong or is there a problem with g_cluster ?

(using Gromacs 4.07)

Many thanks
Andreas