A high-throughput phenotyping assay for precisely determining stalk crushing strength in large-scale sugarcane germplasm

Abstract

Sugarcane is a major industrial crop around the world. Lodging due to weak mechanical strength is one of the main problems leading to huge yield losses in sugarcane. However, due to the lack of high efficiency phenotyping methods for stalk mechanical strength characterization, genetic approaches for lodging-resistant improvement are severely restricted. This study attempted to apply near-infrared spectroscopy high-throughput assays for the first time to estimate the crushing strength of sugarcane stalks. A total of 335 sugarcane samples with huge variation in stalk crushing strength were collected for online NIRS modeling. A comprehensive analysis demonstrated that the calibration and validation sets were comparable. By applying a modified partial least squares method, we obtained high-performance equations that had large coefficients of determination (R2 > 0.80) and high ratio performance deviations (RPD > 2.4). Particularly, when the calibration and external validation sets combined for an integrative modeling, we obtained the final equation with a coefficient of determination (R2) and ratio performance deviation (RPD) above 0.9 and 3.0, respectively, demonstrating excellent prediction capacity. Additionally, the obtained model was applied for characterization of stalk crushing strength in large-scale sugarcane germplasm. In a three-year study, the genetic characteristics of stalk crushing strength were found to remain stable, and the optimal sugarcane genotypes were screened out consistently. In conclusion, this study offers a feasible option for a high-throughput analysis of sugarcane mechanical strength, which can be used for the breeding of lodging resistant sugarcane and beyond.

Plant phenotyping relevance

サトウキビ茎の破砕強度という植物形質を、近赤外分光法と高スループット予測モデルで測定・検証した方法開発および大規模適用が中心である。

Code and data availability

The article describes NIRS phenotyping of sugarcane stalk crushing strength, but no public dataset, code, model, or image repository is provided. The data availability statement only offers raw data upon request, and the supplementary material link is generic with no stated contents of paper-specific data or models.

No evidence-backed public reproduction asset is currently recorded.