Genomic Selection for Nitrogen Use Efficiency Using Canopy Spectral Reflectance in Hard Winter Wheat

The need to breed nitrogen use efficient (NUE) crops is growing due to increasing nitrogen costs and environmental concerns. However, traditional phenotyping methods for NUE are labor intensive and destructive. Canopy spectral reflectance (CSR) can be used as a proxy for physical sampling. During the 2012 growing season, the USDA-NIFA Triticeace Coordinated Agricultural Project (TCAP) supported proximally based CSR phenotyping in the 299-genotype hard winter wheat association mapping panel grown near Ithaca, NE. CSR data was collected once a week for a five week period during grain fill using a dual-fiber optic system with two inter-calibrated Ocean Optics USB 2000+ spectrometers allowing for adjustment to incident light.  This system has 0.4 nm spectral resolution from 350.02 to 1011 nm.  Vegetation indices measuring chlorophyll content and biomass such as NDVI, CI, and EVI were used to investigate plant productivity parameters associated with NUE.  Of the vegetation indices investigated, Enhanced Vegetation Index (EVI) showed the best correlation with biomass, grain yield, and grain N quantity.  A G-BLUP prediction model using SNP markers was applied on the weekly EVI and max EVI value for each entry. Prediction accuracy was estimated using 10 fold cross validation replicated 100 times.   Accuracy for EVI phenotypes ranged from 0.38 for week 1 EVI to 0.57 for week 5 EVI. We concluded that genomic selection combined with CSR data was successful in predicting unphenotyped lines within same year. Further investigation will be done to study the prediction accuracy of genomic selection and CSR data across years/environments.