S05 - Session P4 - In search of the signatures of selection during rose breeding of nineteenth century

Authors: Jeremy Clotault *, Nadine Le Nénaon, Mathilde Liorzou, Nathalie Jacquier, Marie-Christine Le Paslier, Aurélie Bérard, Aurélie Chauveau, Elodie Marquand, Annie Chastellier, Fabrice Foucher, Gilles Michel, Alix Pernet, Vanessa Soufflet-Freslon, Tatiana Thouroude, Cristiana Oghina-Pavie, Agnès Grapin

The 19th century is considered as the golden age of rose breeding due to the large increase in the number of cultivars and the new traits introduced into European roses. However, the genes on which this selection is based are still largely unknown. Moreover, this selection occurred in a particular genetic context in which the old European genetic background was gradually replaced by an Asian genetic background. We have studied how breeding impacted the diversity of candidate genes involved in potentially selected traits in the 19th century. A sample of 365 rose cultivars were sequenced for 11 almost full-length candidate genes involved in flowering period ( RoFT and RoKSN ), number of petals ( AGAMOUS ), petal colour ( DFR , F3'H , FLS , ZDS and CCD1 ), plant architecture ( BRC1 ) and scent ( NUDX1-1 and PAAS ). Moreover, a part of 50 random genes was sequenced on the same sample to evaluate the evolution of the neutral nucleotide diversity. A combination of Access-Array Fluidigm PCR and Illumina sequencing was used to obtain these sequences. Overall, both candidate and random genes evolved in the same direction as the genetic background: the alleles of European roses of the late 19th century are closer to the alleles of Asian roses than to the alleles of early European roses. This trend can either be explained by the fact that Chinese alleles were actually selected for example flowering time, new (yellow) or more intense (red) colours or a different fragrance (tea scent), which may have had a global impact on the genome, reinforced by the low number of generations. But it can also reflect the fact that some traits may have been secondary for breeding. Statistical methods to identify selection signatures more intense than introgression as well as functional data on the detected alleles will help to distinguish between these hypotheses.