Leaf to root: soybean nodule traits monitored non-destructively with leaf spectra
image: Graphic representation of the research hypothesis on the correlation between leaf spectral signatures and the underlying nodule traits (number and total weight) under abiotic stresses (salt and drought).
Credit: Plant Phenomics
Soybean root nodules are central to biological nitrogen fixation, but measuring nodule number and weight usually requires destructive harvesting. This limits repeated monitoring and high-throughput phenotyping. In this study, researchers used leaf hyperspectral reflectance and Partial Least Squares Regression (PLSR) models to estimate nodule traits non-destructively across 528 soybean plants, 18 genotypes, two rhizobium strains, and drought or salt stress. The models predicted nodule number with R² = 0.75 and normalized root mean square error (nRMSE) = 6.02%, and nodule weight with R² = 0.53 and nRMSE = 12.38%. The approach also captured stress- and strain-related spectral patterns and enabled real-time, in situ tracking of nodule development in individual plants, offering a route for sustainable soybean breeding.
Soybean (Glycine max) forms symbiotic root nodules with nitrogen-fixing rhizobia, converting atmospheric nitrogen into ammonia and reducing reliance on synthetic fertilizers. Nodule number and weight are key indicators of symbiotic efficiency and yield potential, yet current quantification relies on excavating roots, excising nodules, and weighing them manually. These methods are destructive, labor-intensive, and poorly suited to longitudinal monitoring or high-throughput phenotyping. Drought, salinity, and rhizobium strain identity can strongly alter nodulation, creating complex trait variation that is difficult to track over time. Based on these challenges, in-depth research is needed to develop non-destructive, scalable tools linking aboveground leaf signals to belowground nodule traits.
In a study published (DOI: 10.1016/j.plaph.2026.100203) in Plant Phenomics and Biochemistry in 2026, researchers at The Chinese University of Hong Kong (CUHK), with collaborators at the University of Hong Kong, Nanjing Agricultural University, and the Guangdong Academy of Sciences, tested whether leaf hyperspectral reflectance could predict soybean nodule traits. The team combined leaf spectra with nodule measurements from 528 plants grown under controlled conditions. They developed Partial Least Squares Regression (PLSR) models that estimated nodule number and weight non-destructively and then tested them across rhizobium strains, drought, and salt stress. The work also tracked nodule development in living plants over time.
The study used 528 soybean plants from 18 genotypes, inoculated with two rhizobium strains: Bradyrhizobium diazoefficiens USDA 110 and Sinorhizobium fredii CCBAU45436. Plants were exposed to low salt (50 mM sodium chloride, NaCl), high salt (150 mM NaCl), drought (15% polyethylene glycol 6000, PEG-6000), or control conditions and harvested at 21 days post-inoculation (dpi). Leaf spectra were preprocessed with Savitzky-Golay smoothing, Z-Score standardization, and first-derivative transformation (SG-ZS-FD). The resulting PLSR models predicted nodule number with R² = 0.75 and nRMSE = 6.02%, and nodule weight with R² = 0.53 and nRMSE = 12.38%. Variable Importance in Projection (VIP) analysis identified shared sensitive wavelengths near 422, 535, 696, 720, 1000, 1140, 1325, 1380, and 1895 nm; nodule weight had unique peaks at 470 and 1720 nm. Drought suppressed nodulation more strongly than salt, and salt stress produced distinct near-infrared and shortwave-infrared fingerprints. Correlations with equivalent water thickness (EWT) were positive, whereas chlorophyll fluorescence indices (CF690 and CF740) and the Photochemical Reflectance Index (PRI) were negative, linking nodule traits to carbon-nitrogen exchange, xanthophyll activity, and leaf water status. In five Wm82 plants, the models tracked nodule number with R² = 0.90 and nodule weight with R² = 0.52.
The authors said the study shifts nodule phenotyping from destructive harvesting to repeated, non-destructive measurement. They said leaf spectra appear to capture systemic carbon-nitrogen exchange and water-status signals, allowing individual plants to be followed over time. They added that the approach could help match soybean genotypes with elite rhizobia and accelerate breeding for stress-resilient symbioses. They said the models performed best for nodule number, a stable structural trait, while nodule weight remains more dynamic and sensitive to water status, soil conditions, and strain identity. Field validation and strain-specific calibration remain important next steps.
The method could enable non-destructive screening of soybean-rhizobium pairs, reducing breeding cycles and helping farmers optimize biological nitrogen fixation. By identifying plants with strong nodulation under drought or salinity, breeders could select resilient genotypes without digging up roots. The authors note that field deployment will require scaling from leaf to canopy, accounting for soil heterogeneity, and managing climatic variability. Combining unmanned aerial vehicle (UAV) hyperspectral imaging with radiative transfer models such as PROSAIL, soil sensors, and time-series monitoring could help preserve the physiological links observed in controlled chambers. If validated in the field, this approach could support more sustainable soybean production with lower synthetic nitrogen fertilizer use.
###
References
DOI
Original Source URL
https://doi.org/10.1016/j.plaphe.2026.100203
Funding information
Hong Kong Research Grants Council Area of Excellence Scheme (AoE/M-403/16); Lo Kwee Seong Biomedical Research Fund; CUHK IdeaBooster Fund (IDBF24SC18); CUHK Direct Grant for Research 2024/2025 (Grant No. 4053685); SLS Seed Fund 2024/2025; National Natural Science Foundation of China–Young Scientists Fund Category C (42507069); Natural Science Foundation of Jiangsu Province (BK20231469); Jiangsu Province Key Research and Development Program (BE2023369); Young Elite Scientists Sponsorship Program by CAST (YESS).
About Plant Phenomics
Plant Phenomics publishes breakthrough research that integrates genomics, genetics, physiology, molecular biology, bioinformatics, statistics, mathematics and computer science to advance plant phenotyping—from the cellular level to entire populations. The journal tackles key scientific challenges in phenomics by showcasing cutting‑edge technologies for data acquisition, management, interpretation and modelling. Its scope spans practical applications across all plant sciences, aiming to translate novel phenotyping tools into real‑world solutions for breeding, crop management and fundamental plant biology. By fostering interdisciplinary innovation, Plant Phenomics serves as a hub for researchers developing high‑throughput, accurate and scalable methods to decipher complex plant traits, ultimately accelerating progress in agriculture and plant science.