Shows SARS-CoV-2’s genome has an extreme CpG-dinucleotide deficiency, more pronounced than in most other coronaviruses, and argues this fits a model of adaptive evasion of the host zinc-finger antiviral protein (ZAP), which preferentially binds and degrades CpG-rich viral RNA. Frames the genome’s dinucleotide/codon-usage composition as a signature consistent with an evolutionary passage history in a mammalian host. relevance_note: the primary codon-usage/genome-composition data point underlying the “the genome carries a natural evolutionary signature” side of the engineering-signal debate.