Predicting structure of protein after a few mutations

Hi all,

I am attempting to predict the structure of a protein after mutating a couple residues (4-5). I have the wildtype structure so I have been adding the mutations one-by-one in pymol, with each single mutation relaxed with the relax function. Is this the best way to use Rosetta to predict such structures? Also, are there any modifications to the relax function that would increase the accuracy for structure prediction?
 

Thank you!