A near-complete genome assembly of onion resolves its genomic architecture, origin, and evolution.

Long-standing controversy over the origin of onion has been compounded by pervasive discordance among phylogenetic inferences, largely due to the absence of a fully resolved reference genome and genome-wide population analyses. Here, we present DH066619-V2, a near-complete genome assembly of onion (Allium cepa) spanning 15.86 Gb, providing unprecedented contiguity for resolving complex repeat architectures. Fine-scale reconstruction of subtelomeric satellite arrays reveals pronounced structural diversity and uncovers a massive, previously overlooked accumulation of large retrotransposon derivatives (LARDs), shedding new light on the mechanisms driving onion genome expansion and lineage-specific evolutionary trajectories within Allium. Population analyses of 122 accessions reconcile conflicting phylogenetic inferences regarding onion origin and trace a photoperiod-driven dispersal history during onion diversification, providing strong genomic support for Allium vavilovii as the wild progenitor. Focusing on S-cytoplasmic male sterility (S-CMS), a cornerstone trait in modern onion hybrid breeding, we delineate the physical interval of the Ms locus and propose that it was introduced into cultivated onion via alien introgression, with Allium roylei as a likely donor of both the S-CMS cytoplasm and the Ms locus. Overall, these results provide a valuable genomic resource for understanding onion genome evolution and facilitate future genetic improvement.