AUTHOR=Patzer Laurine , Schulz Dietmar Frank , Ezeoke Amarachi Queendaline , Linde Marcus , Debener Thomas 

TITLE=Beyond bloom: validated marker–trait discovery for polyploid roses via GWAS

JOURNAL=Frontiers in Plant Science

VOLUME=Volume 16 - 2025

YEAR=2025

URL=https://www.frontiersin.org/journals/plant-science/articles/10.3389/fpls.2025.1591861

DOI=10.3389/fpls.2025.1591861

ISSN=1664-462X

ABSTRACT=IntroductionBreeding roses with ideal ornamental characteristics such as beautiful flowers, a pleasant fragrance, and attractive growth habits is complex and time-consuming. This process can be improved and accelerated through the use of molecular markers.MethodsWe conducted a genome-wide association study  on nine ornamental traits in roses using the RhWagSNP chip across a panel of 285 cultivars and varieties. Significant marker-trait associations found for major quantitative trait loci were further validated using single-marker analyses with PACE technology in independent panels of up to 182 genotypes.ResultsFor six traits— ‘Young shoot: intensity of anthocyanin coloration’, ‘Stem: number of prickles’, ‘Leaf: glossiness of upper side’, ‘Flower: number of petals’, ‘Flower: fragrance’, and ‘Petal: length’—we identified and validated marker-trait associations for major QTLs. Conversely, we were unable to validate associations for ‘Leaf: anthocyanin coloration’ and ‘Leaf: intensity of green color on the upper side’, and found no significant associations in the GWAS for ‘Leaf: size’. Loci that affect petal size, petal number and fragrance have been previously studied. We were able to detect associated markers with increased effect sizes for all three traits. Even greater effects were observed when we combined markers from independent loci for petal number and fragrance.DiscussionAssociated markers for some of the analysed traits largely colocalise with markers previously identified in QTL analyses of biparental populations. Our validation strategy using PACE as an alternative marker technology in independent panels and different environments supports the robustness of our data, irrespective of our limited panel sizes. For the six traits for which we could validate marker-trait associations, our data can be interpreted cautiously as indicating high complexity of inheritance, with few large-effect QTLs influencing the traits. For the other four traits, either greater genetic complexity and/or stronger environmental effects may have confounded our analyses. We believe that the markers presented here can serve as valuable tools for marker-assisted selection and for further genetic analysis of the traits we have analysed.