Genome-wide detection of selection signatures in roses

Modern roses are the result of extensive hybridization among wild species and cultivated forms, followed by strong artificial selection during the past two centuries. Understanding how such selection has shaped the rose genome can provide valuable insight into the genetic basis of key ornamental traits and guide future breeding strategies. To identify genomic regions affected by selection, we analyzed genome-wide single-nucleotide polymorphism (SNP) data across a diverse panel of cut and garden rose cultivars. Observed heterozygosity was calculated for sliding windows of 100 SNPs using 1,000 bootstrap replicates per group, complemented by analyses using physical window definitions. Regions showing significant local reductions in heterozygosity were considered putative selective sweeps. Gene models within these regions were functionally annotated using the Rosaceae Genome Database (Rosa chinensis genome v1.0), integrating Gene Ontology (GO) information. Genome-wide selection scans across the complete rose panel revealed 16 genomic regions exhibiting pronounced reductions in heterozygosity, indicating loci under strong selection during rose domestication. Several of these intervals colocalize with previously characterized genes, most notably RoKSN on chromosome 3, a central regulator of recurrent flowering, and RhPMP1 on chromosome 1, which encodes a plasma membrane protein involved in flower opening. When analyzed separately, cut and garden roses displayed distinct as well as overlapping selective signatures. Cut roses showed pronounced selection signals on chromosomes 3, 6 and 7, overlapping with loci previously associated with floral longevity, dehydration tolerance, and senescence regulation, including RhFer1, RhABF2 and RhNAC2. These associations suggest that the observed selective signals may relate to postharvest performance and stress resilience, although direct phenotypic data for our accessions are not available. Furthermore, a selective sweep region on chromosome 3 colocalized with RoTTG2, a gene associated with prickle density. The concordance with well-characterized loci demonstrates how past breeding has shaped major ornamental traits and points to additional sweep regions that could contain previously uncharacterized candidate genes.