Optimization of evaluation method for low nitrogen tolerance in soybean germplasm seedlings

Abstract

Nitrogen is a critical macro-nutrient for growth and development of soybeans (Glycine max L.). Improving nitrogen use efficiency and developing low nitrogen tolerance varieties are important approaches to mitigate excessive fertilization and maximize production benefits. Precise identification of low nitrogen tolerance germplasms serves as a crucial bridge for converting germplasm advantages into breeding advantages. In this study, we optimized a precise evaluation method for low-nitrogen tolerance in soybean seedlings based on Extreme Gradient Boosting (XGBoost) algorithm. Three hundred soybean germplasms were assessed for low-nitrogen tolerance under hydroponic conditions with normal (7.5 mM) and low (0.75 mM) nitrogen levels. Fourteen physiological traits related to low nitrogen tolerance, such as biomass, chlorophyll fluorescence, were measured. The XGBoost-based evaluation method was compared to a traditional fuzzy membership function comprehensive evaluation method for accuracy and applicability. Results showed that the XGBoost-based method ensured precision and reduced the number of determined physiological indicators compared to traditional methods. Furthermore, this approach reduces the number of traits required for precise identification, which reduces time and improves economic benefits. Consequently, the screening efficiency of soybean low nitrogen tolerance germplasms is improved, offering valuable insights for soybean breeding programs.