The grid was placed into a tank just large enough to hold the grid, and 20 L of hydroponic solution was used per tank. Genotypes were randomly assigned to positions in each cone-tainer grid using the sample and matrix functions in R . Aeration was supplied to each tank with an aquarium pump and two large airstones . Lines were transplanted into the tanks, and the lights and aeration were switched on 24 h after transplanting. When leaf 1 emerged, ¼ strength Hoagland’s solution was added, and the pH was adjusted to 6.5. When leaf 2 emerged, the Hoagland’s was increased to ½ strength in all tanks and CaCl2 was added to the tanks destined for salt treatment in a 15:1 molar ratio of NaCl to CaCl2. When leaf 4 emerged, salt was added to the salt tanks gradually over 2 days to reach a final concentration of 200 mM. The water level and pH were adjusted 3 times per week throughout the experiment. To compare the salt tolerance of four wheat lines,drainage pot both aerial and root biomass was harvested separately for each individual plant 2 weeks after salt treatment was applied. Plant matter was dried at 60-65◦C for 4 days and then weighed.
For linear mixed-effect modeling analysis, dry weight was considered as the response in the analysis, salt level and genotypes were considered as fixed effects, and tank number was considered as a random effect. Model results were identical if tank position was considered as a nested-random effect of tank number, thus the results with tank number were used as the only random effect. To compare the response of the alloplasmics to the response of the euplasmic parents when the salt level is changed, the coefficient estimates of the lme model were examined. All tanks were transferred from the Conviron to greenhouse to take hyperspectral images under natural light conditions. To ensure that each hyperspectral image contained both salt and control plants of a single wheat line, individual cone-tainers were removed from the randomized grid and arranged as salt and control tanks, each containing 12 cone-tainers as shown in Figure 1. After imaging, plants were placed back into their original randomized grid positions to avoid confounding effects from changing the tank position during the experiment. Image acquisition was done ∼24 h after salt application when there were no visual symptoms. To reduce the effects of sun angle and shade, images were captured close to noon . A push broom hyperspectral camera was used for image acquisition, which required constant movement during image capture for two-dimensional spatial information to be accurate.
A glide gear slider was used to mount the camera on a horizontal bar. A Dayton DC gearmotor was utilized to move the slider along at a set speed, with the camera oriented to face downwards. All of this was done as per Moghimi et al. . The camera scanned over 240 spectral channels ranging from 400 to 900 nm with a spectral resolution of about 2.1 nm and captured 640 pixels in the cross track direction . The number of pixels in the along track direction was set to 2,000 to assure both control and salt tanks of each line were captured in a single image. Therefore, the pixel size of each hyperspectral image, also known as hyperspectral data cube, was 2,000 × 640 × 240, meaning each pixel has a 240 dimensional feature vector. The frame rate of the camera was adjusted based on the field of view, the distance between lens and target, and the speed of the camera motion as described by Moghimi et al. . The field of view of the camera lens was 33 degrees, and the distance between the target and lens was about one meter. The speed of the camera was set to 0.025 m/s, thus the calculated frame rate was 27 frames per second to obtain square pixels . Gain and exposure time were adjusted appropriately based on light conditions to avoid over-exposure while taking advantage of the full dynamic range . It should be noted that those vegetation pixels that could pass the segmentation steps might not be pure pixels because of limiting factors such as leaf angle, leaf curvature, and shadow.
Therefore, to extract the spectral signatures for salt stressed and control plants, the most pure pixels for each class should first be identified among the pixels passed from the segmentation steps. These pure pixels can be considered as end members of the two classes. Each hyperspectral image contained only a single wheat line, but included both salt and control treatments. Consequently, there were only two potential classes and subsequently two respective end members in each image. These end members are the most spectrally pure pixels in the hyperspectral image. The assumption of pure pixels existence can be correct because of the high spatial resolution we attained . Based on the strategy proposed by Winter , end member pixels in a feature space are the vertices of a simplex that has the maximum volume compared to any other simplex formed by other pixels. To elaborate, consider each pixel as a point in a d-dimensional feature space where d is the number of bands. From prior assumption, there could be n end members which are pure pixels in the image. These n end members are the vertices of a -simplex that has the maximum volume in a d-dimensional spectral feature space spanned by all pixels . Several algorithms and techniques for extracting end members based on this idea have been developed with the intention of improving computational time and accuracy . To find the unique set of two end members comprising the vertices of a 1-simplex in this study, successive volume maximization was utilized. SVMAX has a modified objective function in which end members are identified recursively through a successive optimization problem . In each image, SVMAX identified two pixels that were the furthest from each other in the high dimensional feature space, each representing one class: salt and control. The primary objective of this research was to quantitatively rank wheat lines based on salt tolerance using HSI. As a case study, four Triticum aestivum bread wheat lines were selected for assessment of salt tolerance with destructive biomass measurements in parallel with HSI. Many different molecular, physiological, and growth parameters can be used to assay salt tolerance differences between genotypes, including Na+ uptake, the ratio of K/Na+, photosynthetic activity , gene expression , and aerial and root biomass in salt versus control conditions over extended growing periods . To assess salt tolerance, salt treatments were applied for 2 weeks, and then root and aerial biomass were measured on a dry weight basis . According to previous studies, Kharchia is a salt tolerant line since it maintains a stable harvest index and yields well in high salt conditions ,drainage planter pot while CS is a salt-sensitive cultivar . Therefore, the main objective of performing conventional phenotyping was to identify the tolerance of the two unknown additional alloplasmic lines, co and sp. The results of biomass measurements for these two lines were compared with CS since they contain the exact same nuclear background as CS, which allowed for a direct comparison of biomass to CS. The biomass measurements revealed that both CS and sp, unlike co, showed a reduction in both aerial and root biomass after growth in the presence of 200 mM NaCl . The analysis of variance found significant interactions at all levels for aerial and root biomass, including between salt level and genotype .
A closer examination of effect sizes using linear mixed modeling showed significantly less change in aerial and root biomass from 0 to 200 mM in co when compared to CS , indicating that the alloplasmic line co is more salt tolerant than the nuclear donor CS in terms of salt effect on biomass. However, overall growth rate may be impacted in co, as biomass in the absence of salt is less than that of CS. A possible explanation for this observation is that altered nuclear-cytoplasmic communication in this line could lead to improper expression of organellar transcripts involved in stress tolerance, therefore “priming” the alloplasmic for stress and reducing sensitivity to salt stress . The response of the other alloplasmic line [i.e., sp] was not significantly different when compared to CS ; however, it trended toward less change in response to salt compared to CS for aerial biomass. The change in root biomass was almost identical to that of CS. Based on these observations, it can be inferred that sp is slightly more tolerant than CS. Although it is historically known that Kharchia is more tolerant than CS, the result of biomass measurements of Kharchia was also compared with CS to examine if the conventional biomass measurement could capture the difference between these two lines with dissimilar genome backgrounds. Intriguingly, the magnitude of aerial biomass change between the control and treatment in the highly salt-tolerant Kharchia cultivar was not significantly different when compared with salt-sensitive CS . However, similar to sp, the trend was also toward a smaller change in response to salt than CS. This indicates that biomass measurement, although a convenient parameter to measure in a lab environment, may not always reflect the actual field performance in desirable traits such as harvest index or yield. This is consistent with previous results that showed a substantial biomass decrease for Kharchia in the presence of salt, yet also a high relative yield and harvest index . Without the substantial historical knowledge of how Kharchia was derived from Indian landraces adapted to sodic soils , the assessment of salt tolerance with hydroponic screening and biomass measurement for this study may have missed this highly valuable source of germplasm. Based on the results of our conventional salt tolerance and historical knowledge, we can conclude that Kharchia and co are more salt-tolerant than sp and CS. In addition, the time-consuming and laborious process of conducing the conventional biomass measurement for salt tolerance assessment underscored the need for more informative and quantitatively precise screening techniques to rapidly and non-destructively assess salt tolerance, particularly when comparing cultivars with drastically different genetic backgrounds and growth regimes, such as Kharchia and CS that differ in vernalization requirements and photoperiod sensitivity . However, since co and sp have identical nuclear backgrounds to CS but only differ in their organellar genomes, direct comparisons of biomass are more valid.Three methods including area under the NRD curve, MDPA, and posterior probability were utilized in this study to analyze hyperspectral images of wheat lines. In all methods, salt stress treatment of each line was compared to its control treatment because differences in spectral responses of a given line may not necessarily be related to the imposed stress, but rather to differences in inherent characteristics such as having waxy and/or darker leaves. The order of ranking of the examined wheat lines was similar for all of these methods. Kharchia was ranked as the most tolerant line followed by co and sp. CS was identified as the most susceptible line by all three methods. In addition to ranking the wheat lines, more inferences could be made from calculating the posterior probability compared to the other two methods. For instance, it could provide the ability to observe the variations of posterior probability over all similarity bins . This observation can be used to define a threshold of making a decision for classification purposes if the classification of salt and healthy plants is of interest. In this section, results and achievements of this research study are discussed. Our findings revealed that conventional phenotyping methods to identify salt tolerant wheat lines could be replaced by the fast and non-invasive methodology proposed in this study. It was surprising to find that the conventional assessment of salt tolerance with biomass measurement was not consistent with the anecdotal and historical evidence of salt tolerance for the Indian land race Kharchia. However, biomass measurements of Kharchia were indeed consistent with previous studies that documented a significant biomass decrease, yet stable harvest index and yield in response to salt stress .