Whole genome analysis of breed-specific copy number variations in dogs

BACKGROUND: Copy number variation (CNV), including genomic gains or losses of DNA segments ranging from kilobases to megabases, represents a major source of genetic diversity and can substantially alter gene dosage, regulation, and phenotype. Although CNVs have been catalogued across many dog breeds in prior large-scale efforts, breed-specific CNVs, defined as copy-number changes that are highly prevalent within one breed but rare or absent across others, have not been systematically investigated. To address this gap, we analyzed whole-genome sequencing data from 436 dogs representing 105 modern breeds and constructed a high-resolution catalog of breed-specific CNVs. RESULTS: Across the cohort, we identified 307 breed-specific CNVs spanning 54 breeds, including 35 CNVs from 20 breeds that fully encompassed annotated gene bodies. Several findings recapitulated known biology, most notably the Rhodesian Ridgeback-specific duplication fully covering FGF3, FGF4, FGF19, and LTO1, which underlies the dorsal hair ridge phenotype. Beyond known loci, we identified previously unreported gene-spanning duplications, including a 309 kb LCORL duplication in English Bulldog, a 206 kb PPP1R17 duplication in Bouvier des Flandres, and a 490 kb MMUT duplication in Bull Terrier. These highly exclusive CNVs might represent strong candidates for artificial selection and/or breed-defining traits. CONCLUSIONS: This study presents a systematic atlas of breed-specific CNVs across 105 modern dog breeds and highlights multiple large-effect, gene-spanning CNVs with potential biological relevance. The analytical framework and CNV catalog establish a foundation for future genotype-phenotype studies in canines and offer a comparative model for understanding how structural variation shapes complex traits across other species, including humans.