UC Davis

Spinach (Spinacia oleracea L., 2n =12) is primarily produced in California, and Arizona, accounting for 95% of the planted spinach area in the US. The heavy application of nitrogen (N) fertilizer is a common practice for dark-green leaf spinach. However, not all applied N is used by the crop, given its short production period. This leads to excessive N leaching, negatively impacting the environmental system and human water sources. We evaluated 344 spinach accessions on high nitrogen-use efficiency (NUE) associated traits including fresh/dry biomass, chlorophyll concentration, total N concentration, nitrate concentration, and leaf shapes under low N environment. Also, we evaluated the effectiveness of high-throughput phenotyping using machine learning models (Chapter 1), and different genomic prediction strategies (Chapter 2) for cost-effective acceleration of the breeding program. Using vegetation indices and texture measures, we trained machine learning models capable of predicting NUE traits in spinach. We improved the prediction ability of genomic prediction models by incorporating vegetation indices in the multi-trait genomic prediction framework. These methods can assist breeding programs in managing time and resources more efficiently. In chapter 3, we investigated the underlying genetic basis of NUE traits. We identified significant markers that are associated with NUE traits using genotype and variant allele frequency in the genome-wide association study.