Calling a complete set Germline+Somatic Mutations for a Cancer Sample

I need to produce a set of all DNA-level variants for cancer patients to serve as exclusion sites for RNA editing site finding. Currently have have matched Tumor-Normal WES data for these patients. Which of the following would be the most appropriate way to produce these?

1. Unfiltered paired MuTect2
2. HC of NORMAL plus filtered paired MuTect2, or
3. HC of TUMOR plus filtered paired MuTect2?