Multivariate and imaging methods to classify cold tolerance of sugarcane (Saccharum spp. hybrids) cultivars and breeding clones.

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.