Univrsity of Tehran Press Iranian Journal of Field Crop Science 2008-4811 56 4 2025 12 22 The Effect of Salt Stress on Wheat Grain Traits using Advanced Imaging Technology The Effect of Salt Stress on Wheat Grain Traits using Advanced Imaging Technology 53 74 105090 10.22059/ijfcs.2025.395914.655142 FA Mujtaba Sadaat Department of Agronomy and Plant Breeding, Faculty of Agriculture, University of Tehran, Karaj, Iran. 0009-0003-8229-3088 Mohammad Reza Bihamta Department of Agronomy and Plant Breeding, Faculty of Agriculture, University of Tehran, Karaj, Iran. 0000-0003-0614-0963 Valiollah Mohammadi Department of Agronomy and Plant Breeding, Faculty of Agriculture, University of Tehran, Karaj, Iran. 0000-0001-6487-0740 Mehrdad Mahlooji Agriculture and Natural Resources Research Center of Isfahan, Isfahan, Iran. 0000-0003-0281-5388 Journal Article 2025 05 24 Introduction. Wheat (Triticum aestivum L.) remains a cornerstone of global food security, yet its productivity is increasingly threatened by soil salinity, which affects over 20% of irrigated lands worldwide. Salinity induces osmotic stress and ion toxicity, leading to a drastic reduction in grain filling duration and efficiency. While traditional breeding focuses on yield components like thousand-kernel weight (TKW), these metrics provide a "black-box" view of grain development. They fail to capture the nuanced changes in grain architecture—such as thickness, sphericity, and surface roughness—which are critical indicators of the physiological health of the plant during the grain-filling stage. The fundamental "Knowledge Gap" in current wheat research lies in the lack of high-throughput, multi-dimensional phenotyping tools capable of screening large germplasm collections. Manual measurements are prone to error and ignore the 3D geometry of the grain. This study addresses this gap by employing a 3D multi-view digital imaging platform to profile the morphometric responses of a massive panel of 320 wheat genotypes. The significance of this study lies in its ability to transition from simple yield-based selection to "Morpho-digital selection," allowing breeders to identify salt-tolerant genotypes based on the stability of their grain architecture rather than just final mass. The objective was to quantify the impact of progressive salinity on 3D grain traits and to establish a diagnostic model using multivariate statistical approaches to identify elite salt-tolerant germplasm. Materials and Methods. The experimental germplasm consisted of a diverse diversity panel of 320 bread wheat genotypes, comprising 108 modern commercial cultivars and 212 Iranian landraces, representing a wide range of genetic plasticity. The research was conducted at the Kabutarabad Agricultural Research Station, Isfahan, Iran (32°99′N, 51°17′E, 1545 m a.s.l.) during the 2022–2023 season. The field layout followed an alpha-lattice design with two replications. Two distinct irrigation regimes were established: A baseline salinity level (S1= 6 dSm-1), which represents the prevailing soil conditions in arid regions, and an elevated salinity stress level (S2= 10 dSm-1) to simulate severe stress. Upon physiological maturity, grains were harvested and prepared for high-throughput phenotyping. A standardized imaging pipeline was developed using a Canon EOS 250D digital camera (24.1 MP) mounted on a controlled-lighting rig. For each genotype, grains were imaged from three orthogonal perspectives: dorsal, lateral, and vertical. This multi-view approach allowed for the extraction of 3D-like volumetric data. Image processing was performed using a custom Python script utilizing the OpenCV (Open Computer Vision) and SciPy libraries. The algorithm performed automated thresholding, contour detection, and feature extraction. Measured traits included primary dimensions (Length/Feret, Width/Breadth, Thickness), area-based indices (Area_D, Area_L, Area_V), and derived geometric parameters (Volume, Aspect Ratio, Compactness, and Centroid Affinity Index). Statistical rigor was ensured through a multi-tiered analysis: ANOVA was performed to determine the significance of Genotype × Environment (G×E) interactions. Linear Discriminant Analysis (LDA) was utilized for group classification, Exploratory Factor Analysis (EFA) for data reduction, and Path Analysis to model the causal relationships between digital traits and final grain mass (TKW). All analyses were conducted in R and SPSS (v.26). Results and Discussion. The results provided an unprecedented look into the "morphological changes" of wheat grains under severe salinity. ANOVA revealed that salinity stress, genotype, and their interaction (G×I) had a highly significant (p< 0.0001) impact on all digital morphometric traits. On average, severe salinity (10 dSm-1) led to a 31.7% reduction in grain volume, which was more pronounced than the reductions in length (26.3%) and width (22.1%). This indicates that salinity primarily inhibits the lateral and vertical expansion of the endosperm, leading to "shriveled" grains. The reduction in mean grain circumference from 43.43 mm to 32.05 mm underscores the inhibition of cell expansion in the grain coat under osmotic pressure. A pivotal finding was the relationship between 3D traits and yield. Correlation analysis showed that while all dimensions decreased, grain thickness emerged as the most reliable predictor of mass, showing a near-perfect correlation with TKW (r= 0.966, p< 0.01). This suggests that in salinity-stressed environments, the "thickness" of the grain is a proxy for the plant's ability to maintain sink strength. Exploratory Factor Analysis (EFA) successfully reduced the complex dataset into four latent factors—Size, Shape/Symmetry, Surface Roughness, and Compactness—explaining 83.5% of the total phenotypic variance. This categorization proves that salinity does not just make grains smaller; it fundamentally alters their geometric symmetry and surface texture. The Linear Discriminant Analysis (LDA) model demonstrated extraordinary diagnostic power, classifying genotypes into S1 and S2 groups with 96% accuracy. The high discriminant coefficients for Area_D and the Concentricity Index (CAI) suggest that these digital traits can serve as "bio-signatures" for salinity stress. Furthermore, Path Analysis elucidated the internal mechanism of yield loss: grain volume (Vol_V) exerted the largest direct positive effect on TKW (β= 0.95), acting as a central mediator for all other dimensional traits. Interestingly, the negative direct effect of thickness on TKW in the path model (β= –0.11), despite its high raw correlation, suggests a physiological trade-off where the plant sacrifices grain sphericity to maintain some level of mass under stress. Compared to previous studies (e.g., Zhang et al., 2022) which focused on 2D measurements, our 3D multi-view approach captured the "cylindricity" and "volume" more accurately, explaining an additional 15-20% of the variance in grain weight. The identification of the "Ohadi" cultivar (LDA score= 4.678) as a top performer highlights the potential of using Iranian landrace genetics to improve the resilience of modern cultivars. Conclusion. This study revealed that salinity stress disrupts the 3D architecture of the wheat grain in a predictable and quantifiable manner. Grain thickness and volume were identified as the most critical digital biomarkers for salinity tolerance. The 3D multi-view imaging pipeline developed here offers a non-destructive, rapid, and cost-effective alternative to traditional lab-based measurements. We recommend that breeding programs integrate "Area_D" and "Concentricity Index" into their selection indices to screen for genotypes that maintain grain filling integrity under high-salinity environments. Future research could link these digital phenotypes with SNP markers through Genome-Wide Association Studies (GWAS) to uncover the underlying genetic loci. Introduction. Wheat (Triticum aestivum L.) remains a cornerstone of global food security, yet its productivity is increasingly threatened by soil salinity, which affects over 20% of irrigated lands worldwide. Salinity induces osmotic stress and ion toxicity, leading to a drastic reduction in grain filling duration and efficiency. While traditional breeding focuses on yield components like thousand-kernel weight (TKW), these metrics provide a "black-box" view of grain development. They fail to capture the nuanced changes in grain architecture—such as thickness, sphericity, and surface roughness—which are critical indicators of the physiological health of the plant during the grain-filling stage. The fundamental "Knowledge Gap" in current wheat research lies in the lack of high-throughput, multi-dimensional phenotyping tools capable of screening large germplasm collections. Manual measurements are prone to error and ignore the 3D geometry of the grain. This study addresses this gap by employing a 3D multi-view digital imaging platform to profile the morphometric responses of a massive panel of 320 wheat genotypes. The significance of this study lies in its ability to transition from simple yield-based selection to "Morpho-digital selection," allowing breeders to identify salt-tolerant genotypes based on the stability of their grain architecture rather than just final mass. The objective was to quantify the impact of progressive salinity on 3D grain traits and to establish a diagnostic model using multivariate statistical approaches to identify elite salt-tolerant germplasm. Materials and Methods. The experimental germplasm consisted of a diverse diversity panel of 320 bread wheat genotypes, comprising 108 modern commercial cultivars and 212 Iranian landraces, representing a wide range of genetic plasticity. The research was conducted at the Kabutarabad Agricultural Research Station, Isfahan, Iran (32°99′N, 51°17′E, 1545 m a.s.l.) during the 2022–2023 season. The field layout followed an alpha-lattice design with two replications. Two distinct irrigation regimes were established: A baseline salinity level (S1= 6 dSm-1), which represents the prevailing soil conditions in arid regions, and an elevated salinity stress level (S2= 10 dSm-1) to simulate severe stress. Upon physiological maturity, grains were harvested and prepared for high-throughput phenotyping. A standardized imaging pipeline was developed using a Canon EOS 250D digital camera (24.1 MP) mounted on a controlled-lighting rig. For each genotype, grains were imaged from three orthogonal perspectives: dorsal, lateral, and vertical. This multi-view approach allowed for the extraction of 3D-like volumetric data. Image processing was performed using a custom Python script utilizing the OpenCV (Open Computer Vision) and SciPy libraries. The algorithm performed automated thresholding, contour detection, and feature extraction. Measured traits included primary dimensions (Length/Feret, Width/Breadth, Thickness), area-based indices (Area_D, Area_L, Area_V), and derived geometric parameters (Volume, Aspect Ratio, Compactness, and Centroid Affinity Index). Statistical rigor was ensured through a multi-tiered analysis: ANOVA was performed to determine the significance of Genotype × Environment (G×E) interactions. Linear Discriminant Analysis (LDA) was utilized for group classification, Exploratory Factor Analysis (EFA) for data reduction, and Path Analysis to model the causal relationships between digital traits and final grain mass (TKW). All analyses were conducted in R and SPSS (v.26). Results and Discussion. The results provided an unprecedented look into the "morphological changes" of wheat grains under severe salinity. ANOVA revealed that salinity stress, genotype, and their interaction (G×I) had a highly significant (p< 0.0001) impact on all digital morphometric traits. On average, severe salinity (10 dSm-1) led to a 31.7% reduction in grain volume, which was more pronounced than the reductions in length (26.3%) and width (22.1%). This indicates that salinity primarily inhibits the lateral and vertical expansion of the endosperm, leading to "shriveled" grains. The reduction in mean grain circumference from 43.43 mm to 32.05 mm underscores the inhibition of cell expansion in the grain coat under osmotic pressure. A pivotal finding was the relationship between 3D traits and yield. Correlation analysis showed that while all dimensions decreased, grain thickness emerged as the most reliable predictor of mass, showing a near-perfect correlation with TKW (r= 0.966, p< 0.01). This suggests that in salinity-stressed environments, the "thickness" of the grain is a proxy for the plant's ability to maintain sink strength. Exploratory Factor Analysis (EFA) successfully reduced the complex dataset into four latent factors—Size, Shape/Symmetry, Surface Roughness, and Compactness—explaining 83.5% of the total phenotypic variance. This categorization proves that salinity does not just make grains smaller; it fundamentally alters their geometric symmetry and surface texture. The Linear Discriminant Analysis (LDA) model demonstrated extraordinary diagnostic power, classifying genotypes into S1 and S2 groups with 96% accuracy. The high discriminant coefficients for Area_D and the Concentricity Index (CAI) suggest that these digital traits can serve as "bio-signatures" for salinity stress. Furthermore, Path Analysis elucidated the internal mechanism of yield loss: grain volume (Vol_V) exerted the largest direct positive effect on TKW (β= 0.95), acting as a central mediator for all other dimensional traits. Interestingly, the negative direct effect of thickness on TKW in the path model (β= –0.11), despite its high raw correlation, suggests a physiological trade-off where the plant sacrifices grain sphericity to maintain some level of mass under stress. Compared to previous studies (e.g., Zhang et al., 2022) which focused on 2D measurements, our 3D multi-view approach captured the "cylindricity" and "volume" more accurately, explaining an additional 15-20% of the variance in grain weight. The identification of the "Ohadi" cultivar (LDA score= 4.678) as a top performer highlights the potential of using Iranian landrace genetics to improve the resilience of modern cultivars. Conclusion. This study revealed that salinity stress disrupts the 3D architecture of the wheat grain in a predictable and quantifiable manner. Grain thickness and volume were identified as the most critical digital biomarkers for salinity tolerance. The 3D multi-view imaging pipeline developed here offers a non-destructive, rapid, and cost-effective alternative to traditional lab-based measurements. We recommend that breeding programs integrate "Area_D" and "Concentricity Index" into their selection indices to screen for genotypes that maintain grain filling integrity under high-salinity environments. Future research could link these digital phenotypes with SNP markers through Genome-Wide Association Studies (GWAS) to uncover the underlying genetic loci.

https://ijfcs.ut.ac.ir/article_105090_8415e556d9e95d7a24e1ec44bb7dd9fa.pdf