Genomic reconstruction of upland cotton domestication uncovers staged selection, gene flow, and flowering-time adaptation

Upland cotton ( Gossypium hirsutum L.) makes up about 90% of global cotton farming. Despite its importance, the origins and domestication history of upland cotton remain poorly understood. Here, we present a high-density pan-genome variation map constructed from 2,910 cotton accessions, including 440 newly sequenced G. hirsutum landraces. Our pan-genome analysis indicates that modern upland cotton most probably originated from a single domestication and underwent three major stages. We also identified several genomic signatures associated with agriculturally important traits, including photoperiod sensitivity, fiber properties, and seed yield, which map to candidate loci GhTOFD06 , GhFLD11, and GhSID05, respectively, through a genome-wide association study (GWAS) and VIGS knockdown. During the third domestication stage (D3), selective pressures favored genes, such as GhTOFD06, a homolog of Arabidopsis COP9 signalosome complex subunit 5b (CSN5B), that control photoperiod-regulated flowering, thereby accelerating cotton domestication and facilitating upland cotton’s latitudinal expansion. Notably, functional validation via gene silencing showed that knockdown of GhSID05 reduced single seed weight by 11.41%, confirming its pivotal role in seed yield regulation. Additionally, we revealed natural gene flow between G. hirsutum and G. barbadense , which has significantly enriched the genetic diversity of the G. hirsutum gene pool and may have contributed favorable alleles for improving modern upland cotton. Our study provides a comprehensive understanding of the genomic evolution of G. hirsutum and valuable genetic resources for future breeding programs.