Genetic Structure and Selection Signature in Flora Scent of Roses by Whole Genome Re-Sequencing

Roses are important plants, and they are cultivated worldwide for their beautiful flowers, remarkable scent, and nutritious hips. In this study, we re-sequenced the whole genomes of 57 rose cultivars and one wild species that originated from different regions around the world and had different scents, aiming to evaluate their genetic structure and to detect the potential signature of the selective sweep between different scent groups with single-nucleotide polymorphism (SNP) and indel markers. The roses were sequenced at an average depth of 6x to the reference genome of Rosa 'Old blush'. A total of 2,375,949 SNPs were obtained. The SNP numbers varied among the 58 samples, with an average of 1,271,906 per sample. The phylogeny and population structure revealed that the roses could be divided into three main clusters; however, the groups were not consistent in terms of geographic origin or scent classification. These indicated that rose cultivars have a complex genetic background due to the extensive hybridization between cultivated roses or wild rosa species worldwide. A selective sweep analysis was conducted to detect the selection signatures in rose scent traits. A total of 2430 candidate genes were identified in the strong scent groups, and were at the top 5% F-ST compared to the moderate group. These genes were significantly enriched in the KEGG pathways of tyrosine metabolism, cyanoamino acid metabolism, alpha-linolenic acid metabolism (13 genes), and phenylpropanoid biosynthesis (30 genes). When the low-scent group was used as the control, 2604 candidate genes were identified at the top 5% F-ST that were significantly enriched in the KEGG pathways of alpha-linolenic acid metabolism (15 genes), and glutathione metabolism (23 genes). We also observed genes enriched in pathways including the linoleic acid metabolism (five genes), diterpenoid biosynthesis (six genes), and monoterpenoid biosynthesis (seven genes), although they were not significant. These implied a positive selection of rose cultivars with a strong fragrance in terms of fatty acid derivatives, terpenoids and benzenoids/phenylpropanoids during rose breeding. Our study establishes a foundation for the further large-scale characterization of rose germplasm, improving the genetic knowledge of the background of roses.