Karl SchmidHappy to share our preprint on genetic mapping and genomic prediction of Distinctness, Uniformity and Stability (DUS) characteristics in European #maize varieties.
Work led by our postdoc Anurag Daware with 🇪🇺🇩🇪🇫🇷🇭🇺🇦🇹 plant variety examination offices.
https://www.biorxiv.org/content/10.64898/2026.06.10.731330v1
Leveraging genome-wide association studies and genomic prediction for distinctness, uniformity, and stability (DUS) testing in maize
Testing for distinctness, uniformity, and stability (DUS) is a requirement for plant variety registration and based on phenotypic traits, which is time-consuming and sensitive to envi-ronmental variation. Advances in genomics allow to complement DUS testing with molecular markers, for which two models in DUS testing were proposed by the Union for the Protection of New Varieties of Plants (UPOV). A use cases was described for maize, but an implementation has been hindered by a lack of suitable markers and validated analytical frameworks. We ad-dress these challenges by integrating historical DUS characteristics scores from 352 European hybrid maize varieties with high-density genome-wide single nucleotide polymorphism (SNP) data. Using genome-wide association studies (GWAS), we identified 18 genomic regions and candidate genes associated with 12 DUS characteristics, enabling the development of diag-nostic markers consistent with the UPOV model "Characteristic-Specific Molecular Markers". Since most DUS traits are polygenic, we combined GWAS-informed marker selection with XG-Boost-based machine learning to predict notes of DUS characteristics. This approach achieved strong predictive performance across multiple traits (mean accuracy 0.67), demonstrating its potential for managing reference collections under UPOV model "Combining phenotypic and molecular distances in the management of variety collections". Both approaches were validated for two characteristics using independent public USDA-NPGS maize datasets (>1,700 accessions) highlighting the value of public data for method validation. We also identify key limitations of historical DUS data, including imbalanced and sparse trait representation, and discuss mitigation strategies. Despite these constraints, our results demonstrate that molecular markers may improve maize DUS testing, enabling faster, more accurate variety registration and supporting accelerated crop improvement. ### Competing Interest Statement The authors have declared no competing interest. European Unions Horizon 2020 research and innovation programme, 817970
JIPB
DoomsdaysCW
in silico Plants
Joseph Lim 
burque.fun: stuff happening in albuquerque
