A high-throughput phenotyping method for sugarcane rind penetrometer resistance and breaking force characterization by near-infrared spectroscopy

Abstract

Abstract Background Sugarcane ( Saccharum spp. ) is the core crop for sugar and bioethanol production over the world. A major problem in sugarcane production is stalk lodging due to weak mechanical strength. Rind penetrometer resistance (RPR) and breaking force are two kinds of regular parameters for mechanical strength characterization. However, due to the lack of efficient methods for determining RPR and breaking force in sugarcane, genetic approaches for improving these traits are generally limited. This study was designed to use near-infrared spectroscopy (NIRS) calibration assay to accurately assess mechanical strength on a high-throughput basis for the first time. Results Based on well-established laboratory measurements of sugarcane stalk internodes collected in the years 2019 and 2020, considerable variations in RPR and breaking force were observed in the stalk internodes. Following a standard NIRS calibration process, two online models were obtained with a high coefficient of determination ( R 2 ) and the ratio of prediction to deviation (RPD) values during calibration, internal cross-validation, and external validation. Remarkably, the equation for RPR exhibited R 2 and RPD values as high as 0.997 and 17.70, as well as showing relatively low root mean square error values at 0.44 N mm −2 during global modeling, demonstrating excellent predictive performance. Conclusions This study delivered a successful attempt for rapid and precise prediction of rind penetrometer resistance and breaking force in sugarcane stalk by NIRS assay. These established models can be used to improve phenotyping jobs for sugarcane germplasm on a large scale.

Plant phenotyping relevance

サトウキビ茎の機械的強度という植物形質をNIRSで高スループット推定する手法を開発・検証しており、表現型取得が研究の中心である。

Code and data availability

The article describes NIRS models for sugarcane RPR and breaking force, but the only availability statement says all data are included in the article/supplementary files, with no public repository, deposit, or authors' URL for datasets, spectra, models, or analysis code. No paper-specific public asset is actionable.

No evidence-backed public reproduction asset is currently recorded.