Pangenome graphs reveal the extent and complexity of genetic variation in North America's most abundant mammal

Pangenome methods reduce reference bias and enable the characterization of previously inaccessible genomic variation within a species, and are thus especially useful for species with high levels of genetic diversity. Here, we constructed a pangenome for the deer mouse (Peromyscus maniculatus), a model for studying local adaptation and a potent reservoir for diverse zoonotic diseases, using long-read assemblies from 14 ecologically and biogeographically diverse populations and 2 outgroup species (totaling 19 haplotypes). The total length of the pangenome is 3.9 Gb, but only 1.4 Gb is shared across all haplotypes, suggesting vast structural diversity. From this pangenome we identified ~108.3 million single nucleotide polymorphisms (SNPs), ~23.9 million indels (variants 50 bp) as well as complex genetic variation driven by massive chromosomal rearrangements and an unprecedented diversity of centromeric satellite positions. Furthermore, we uncover widespread gene copy number variation (CNVs) -some genes vary in number by an order of magnitude across deer mouse haplotypes- in addition to several gene-rich loci that display high levels of structural diversity. Such genes are enriched for functions related to sensory perception, environmental interaction, and immune responses and many display signatures of positive selection at the molecular level, suggesting functional diversification. Our work provides important insight into the utility of pangenomes for genetically diverse species and illuminates underappreciated genomic plasticity within a mammalian species.