Pedigree-reconciled pan-genomics reveals megabase-scale hitchhiking after a century of canola breeding

Abstract

Canola breeding has been shaped by strong selection for oil quality, yet the origins of the known oil quality alleles and genomic consequences of their selection are not fully resolved. By integrating pedigree reconstruction with graph pan-genomics we trace inheritance of ancestral genomic regions across historical and contemporary germplasm. Surrounding the low erucic acid allele in BnA08.FAE1, we identify a 17.23 Mb haplotype that approached fixation in Australian canola in the early 2000s. Contradicting the prevailing model, this haplotype predates modern breeding and was likely widespread in ancestral B. napus in the early 1900s. Genomic analyses implicate centromeric recombination suppression and structural variation in its long-term persistence, which has led to megabase-scale diversity loss through hitchhiking of neighbouring alleles. The haplotype contains extensive structural variation and multiple alleles associated with polygenic disease resistance. Together, these findings reveal the long-term consequences of repeated selection on standing variation during crop improvement.