Unverified paper record
Multivariate and imaging methods to classify cold tolerance of sugarcane ( Saccharum spp. hybrids) cultivars and breeding clones.
Frontiers in plant science · 7 Aug 2026 · 10.3389/fpls.2026.1882707
Abstract
Tolerance against winter freeze is the main focus of variety development in Louisiana, which represents the northernmost sugarcane-growing region worldwide. Antifreeze metabolites, xylem structure, and fiber content represent interrelated physicochemical properties contributing to freeze tolerance. This study first classified the cold tolerance of sugarcane cultivars using metabolites in juice as predictor variables. The best-fit model (XGBoost discriminant analysis) estimated the higher cold tolerance of the final on-station clone progeny to the stress tolerance-inducing wild germplasm line. Stalks of the tolerant sugarcane genotype contained higher fiber for mechanical support against cellular injury, compared to susceptible varieties. Fluorescence microscopy visualized phospholipids responsible for maintaining membrane fluidity during frost in lignin surrounding the vascular bundle. Thermal imaging is proposed for real-time monitoring of spatiotemporal temperature changes, as stalk injury is initiated by ice formation at sub-freeze temperatures during winter freeze. As additional datasets for independent prediction become available, developed methods could be used to explore the biomarkers for stress resistance in simpler multivariate discriminant analysis and the distribution of specific biomarkers in cellular components by microscopic imaging, and to trace stalk injury hot spots as a function of time and relationships with resistance markers.
Plant phenotyping relevance
サトウキビの耐寒性という植物状態を、XGBoost判別モデルと画像・熱画像によって分類・評価する方法が研究の中心であり、単なる生物学的測定ではない。
titleMultivariate and imaging methods to classify cold tolerance of sugarcane ( Saccharum spp. hybrids) cultivars and breeding clones.
abstractThis study first classified the cold tolerance of sugarcane cultivars using metabolites in juice as predictor variables. The best-fit model (XGBoost discriminant analysis) estimated the higher cold tolerance of the final on-station clone progeny to the stress tolerance-inducing wild germplasm line.
abstractThermal imaging is proposed for real-time monitoring of spatiotemporal temperature changes, as stalk injury is initiated by ice formation at sub-freeze temperatures during winter freeze.
Code and data availability
The article blocks contain no public phenotype dataset, image/sensor inputs, author code repository, or trained model deposit. The data availability statement only points to the article/Supplementary Material (Table S1 is a list of varieties/parents, not the phenotyping measurements) and directs further inquiries tothe
No evidence-backed public reproduction asset is currently recorded.
This is an automatically classified, unverified record. Curator approval is required before any resource enters the Catalog.