Using Spectral Reflectance for Genome Wide Association Study and Germplasm Selection for Yield in Pacific Northwest Winter Wheat Breeding Programs.

Spectral reflectance technology has recently opened up new possibilities to characterize traits that are resource-intensive or difficult to measure directly in large germplasm collections, especially breeding programs. In previous studies, we identified vegetation indices that have selectable genetic variations, strong associations with yield, and moderate to high efficiency of indirect selection in Pacific Northwest winter wheat. In this study, we performed genome-wide association study (GWAS) for these selected indices with a prime objective of assessing their potential use to identify genomic regions of agronomic importance. Eight vegetation indices were derived from the reflectance value: green normalized difference vegetation index (GNDVI), normalized chlorophyll-pigment ratio index (NCPI), normalized difference vegetation index (NDVI) normalized water index (NWI), plant nitrogen content (PNC), photochemical reflectance index (PRI), and simple ratio (SR). The indices were considered as traits in genome-wide association study that uses 3,653 SNP markers, two population subgroups and reduced kinship matrices. Comparative analysis of positional proximity and pleotropic effects was done with respect to the QTL regions identified for agronomic traits in the preceding experiment of this project and in other reports. We identified 80 QTL for these spectral vegetation indices across 16 chromosomes most of which were closely localized with grain yield, grain number per spike, thousand kernel weight, test weight, plant height, and heading date. This study demonstrates the impending possibility of using canopy spectral reflectance in identifying novel as well as previously known QTL for yield and component traits in variable environments. Selection of germplasm in breeding programs with enhanced stress tolerance can be done either through canopy spectral reflectance or through marker-assisted selection.