Relaxation vs minimisation

Hi all,

I am very new to the computational structural biology community and I have tried to model a structure by using a software which runs MODELLER on the background. However, my result shows a number of steric clashes and a very high fa_rep when I calculate it with PyRosetta. I am therefore trying to improve the structure before moving on with the rest of my analysis. 

I have so far implemented a script that uses MinMover() to minimise the structure, with very little success, and I am now in the process of implementing FastRelax. However, I don't quite grasp conceptually what the difference between minimising and relaxing a structure is? Am I correct to think that the MinMover only allows the residues to move about their actual position, whilst FastRelax "heats up" the structure to then cool it down to a more favourable energy, whilst still trying to stick as close as possible to the original structure?

Also, should I be running these two multiple times on the same structure? My understanding is that the MinMover is already iterative and only stops when it reaches the threshold I set?

On a related but side note, if I then wanted to compare my modelled structure with PDB structures, should I be minimising those too first? I know that crystals might constrain the "real" structure and pack it tighter, but there are no obvious steric clashes in the PDB structures I have looked at so far.

Thank you for your help!

M.