[modeller_usage] query

Hi, i wish to model a protein having known pdb structure. After getting a model from modeller and evaluating it I found the Dope scores from template_evaluation.py (-279092.718750)and from model_evaluation.py (-26497.847656). How do I refine the model further so that it becomes comparable with the known pdb structure? Why do I get a violatins file,Does this mean that my model is incorrect? The pdb structure is in form of the protein-NAD adduct, how do I insert NAD in my model. Can I separately model the active site and incorporate a ligand(not biological)in my model?

Thanks and regards, Deepti