I have a question about using MODELLER. I've built models of a protein that I'm currently studying and wanted to know if it's necessary to minimize regions of the protein that may have poor quality based on the QMEAN score. Or does MODELLER minimize these regions when building homology models?