Gains across entire genome?

[j-gillespie-dna]

Caveat: I am using the nf-core/sarek pipeline to run CNVkit

I have paired T/N whole exome data. I made a reference with normal samples. However, when I then use that reference to call cnv in samples, both normal and tumor show gains across the entire genome. When I open the diagram pdf, almost every chromosome is red, showing a gain. I find this strange as these are the normal samples used to make the reference.

Has anyone seen this before? Is there some setting/variable I am missing?