Digital phenotyping reveals waterlogging-tolerant chrysanthemum varieties

The approach not only reduces the need for labor-intensive traditional evaluations but also uncovers genetic resources that could safeguard future breeding programs.

Between 2006 and 2016, flooding accounted for nearly two-thirds of global crop losses. Waterlogging creates oxygen-deficient soils, leading to yellowing, wilting, and ultimately plant death. Chrysanthemums, a cornerstone of the global flower industry, are highly vulnerable to waterlogging stress that reduces quality, yield, and survival. Traditional evaluations of waterlogging tolerance rely on manual measurements—plant height, chlorosis scores, and biomass—that are destructive, slow, and subjective. The urgent need for reliable and high-throughput methods has driven interest in nondestructive imaging technologies. Among these, red-green-blue (RGB) imaging offers affordability and ease of use, making it an attractive tool for rapid stress phenotyping.

A study (DOI: 10.1016/j.plaphe.2025.100019) published in Plant Phenomics on 6 March 2025 by Fadi Chen’s team, Nanjing Agricultural University, provides a fast, low-cost, and highly effective strategy for overcoming the phenotyping bottleneck in chrysanthemum breeding.

In this study, researchers employed both traditional trait measurement and advanced digital imaging methods to evaluate the waterlogging tolerance of 180 cultivated and wild chrysanthemum accessions at two stress stages: moderate (T1) and severe (T2). Traditional phenotyping, which recorded traits such as leaf yellowing, wilting, and biomass, revealed significant variability in tolerance, with plants grouped into five grades from highly tolerant to susceptible. Most accessions fell into the “less tolerant” category, while highly tolerant resources were concentrated among wild relatives and intergeneric hybrids, particularly Artemisia. To complement these manual assessments, the team captured over 43,000 RGB images per plant from multiple angles, generating 103 image-based traits (i-traits) including morphological and texture features. Correlation analyses confirmed that i-traits, such as image-based plant height and projected green area, closely matched manual measurements of height and fresh weight, underscoring their reliability as nondestructive indicators. Variation across i-traits was substantial, with many showing high heritability, suggesting strong genetic control over waterlogging responses. Further analysis identified key morphological i-traits, such as convex hull area and total projected area, which effectively reflected biomass changes under stress. Additionally, novel texture-based traits derived from hue and value components were highly correlated with traditional indices, making them powerful early markers of waterlogging tolerance. Machine learning models, particularly random forest, were then applied to predict aboveground biomass and membership function values of waterlogging tolerance (MFVW) using i-trait datasets. These models achieved strong performance, with R² values exceeding 0.85 in some cases. Importantly, cross-validation showed that a reduced set of just 13 top-ranked i-traits could maintain high prediction accuracy, streamlining the evaluation process. Together, these results demonstrate that integrating RGB imaging and machine learning offers a rapid, accurate, and cost-effective strategy for quantifying waterlogging tolerance and identifying elite genetic resources for chrysanthemum breeding.

The integration of digital imaging and machine learning into plant breeding pipelines holds great promise for floriculture and agriculture. By streamlining the identification of tolerant germplasm, breeders can accelerate the development of resilient cultivars, reducing production losses under increasingly frequent flood events. The approach also lowers costs for breeders by eliminating the need for destructive biomass measurements and subjective chlorosis scoring, making advanced phenotyping accessible even to smaller breeding programs. Beyond chrysanthemums, this framework could be adapted for other ornamental and crop species facing similar stress challenges, contributing to global food and flower security.

###

References

DOI

10.1016/j.plaphe.2025.100019

Original URL

https://doi.org/10.1016/j.plaphe.2025.100019

Funding information

This research was funded by the National Key Research and Development Program of China (2023YFD2300900), the National Natural Science Foundation of China (32102421, 32271938), the Program for Key Research and Development Jiangsu, China (BE2023367), the Hainan Provincial Natural Science Foundation of China (323CXTD386), the JBGS Project of Seed Industry Revitalization in Jiangsu Province (JBGS [2021]020), the China Agriculture Research System (CARS-23-A18), the Fundamental Research Funds for the Central Universities (QTPY202005), and a project funded by the Priority Academic Program Development of Jiangsu Higher Education Institutions.

About Plant Phenomics

Science Partner Journal Plant Phenomics is an online-only Open Access journal published in affiliation with the State Key Laboratory of Crop Genetics & Germplasm Enhancement, Nanjing Agricultural University (NAU) and distributed by the American Association for the Advancement of Science (AAAS). Like all partners participating in the Science Partner Journal program, Plant Phenomics is editorially independent from the Science family of journals. Editorial decisions and scientific activities pursued by the journal's Editorial Board are made independently, based on scientific merit and adhering to the highest standards for accurate and ethical promotion of science. These decisions and activities are in no way influenced by the financial support of NAU, NAU administration, or any other institutions and sponsors. The Editorial Board is solely responsible for all content published in the journal. To learn more about the Science Partner Journal program, visit the SPJ program homepage.