We designed an image-based phenotyping protocol to determine the morphological and physiological responses to single and combined heat, drought, and waterlogging treatments. This approach enabled the identification of early, late, and recovery responses at a whole plant level, particularly above-ground parts, and highlighted the necessity of using multiple imaging sensors.
High throughput image-based phenotyping is a powerful tool to non-invasively determine the development and performance of plants under specific conditions over time. By using multiple imaging sensors, many traits of interest can be assessed, including plant biomass, photosynthetic efficiency, canopy temperature, and leaf reflectance indices. Plants are frequently exposed to multiple stresses under field conditions where severe heat waves, flooding, and drought events seriously threaten crop productivity. When stresses coincide, resulting effects on plants can be distinct due to synergistic or antagonistic interactions. To elucidate how potato plants respond to single and combined stresses that resemble naturally occurring stress scenarios, five different treatments were imposed on a selected potato cultivar (Solanum tuberosum L., cv. Lady Rosetta) at the onset of tuberization, i.e. control, drought, heat, waterlogging, and combinations of heat, drought, and waterlogging stresses. Our analysis shows that waterlogging stress had the most detrimental effect on plant performance, leading to fast and drastic physiological responses related to stomatal closure, including a reduction in the quantum yield and efficiency of photosystem II and an increase in canopy temperature and water index. Under heat and combined stress treatments, the relative growth rate was reduced in the early phase of stress. Under drought and combined stresses, plant volume and photosynthetic performance dropped with an increased temperature and stomata closure in the late phase of stress. The combination of optimized stress treatment under defined environmental conditions together with selected phenotyping protocols allowed to reveal the dynamics of morphological and physiological responses to single and combined stresses. Here, a useful tool is presented for plant researchers looking to identify plant traits indicative of resilience to several climate change-related stresses.
The potential effects of climate change, including the increase in the intensity and frequency of heat waves, flooding, and drought events, have negative impacts on growing crops1. It is important to understand the influence of climate change on crop variability and the consequent fluctuations in annual crop production2. With increasing population and food demand, maintaining the yield of crop plants is a challenge, thereby, finding climate-resilient crops for breeding is urgently required3,4. Potato (Solanum tuberosum L.) is one of the essential food crops that contributes to global food security because of its high nutritional value and increased water use efficiency. However, reduction in growth and yield under unfavorable conditions is a main problem, particularly in the susceptible varieties5,6. Many studies highlighted the importance of investigating alternative approaches to maintain potato crop productivity, including agricultural practices, finding tolerant genotypes, and understanding the impact of stress on the development and yield7,8,9, which is also highly demanded by European potato growers (or farmers)10.
Automated phenotyping platforms, including image-based phenotyping, enable the quantitative analyses of plant structure and function that are essential for selecting relevant traits of interest11,12. High throughput phenotyping is an advanced non-invasive technique to determine various morphological and physiological traits of interest in a reproducible and rapid manner 13. Although phenotype reflects genotypic differences in connection to environmental effects, comparing plants under controlled conditions with stress enables linking the extensive phenotyping information to a specific (stress) condition14. Image-based phenotyping is essential in describing phenotypic variability, and it is also capable of screening a set of traits across plant development regardless of the population size15. For instance, the measurement of morphological traits, including the shape, size, and color index of leaves using Red-Green-Blue (RGB) imaging sensors, is used to determine plant growth and development. Moreover, measurements of physiological traits, including photosynthetic performance, canopy temperature, and leaf reflectance, are quantified using multiple types of sensors, such as chlorophyll fluorescence, thermal infrared (IR), and hyperspectral imaging16. Recent studies in controlled environments showed the potential of using image-based phenotyping in assessing different mechanisms and physiological responses of plants under abiotic stresses such as heat in potato17, drought in barley18, rice19, and combined drought and heat treatments in wheat20. Even though studying the responses of plants to multiple stress interactions is complex, the findings reveal new insights in understanding plant mechanisms in coping with rapid change in climate conditions21.
Plant physiological and morphological responses are directly influenced by abiotic stress conditions (high temperature, water deficit, and flooding), resulting in yield reduction22. Even though potatoes have a high water use efficiency compared to other crops, water deficit negatively affects the yield quantity and quality due to the shallow root architecture5. Depending on the intensity and duration of drought level, the leaf area index is reduced, and retardation in canopy growth with inhibition of new leaf formation is pronounced during later stages of stress leading to a reduction in the photosynthetic rate23. The threshold level of water is critical with excess water or prolonged drought periods, resulting in a negative effect on plant growth and tuber development due to oxygen limitation, decreased root hydraulic conductivity, and restriction of gas exchange24,25. Moreover, potatoes are sensitive to high temperatures where temperatures above optimum levels result in delayed tuber initiation, growth, and assimilation rates26. When stresses appear in combination, the biochemical regulations and physiological responses differ from individual stress responses, highlighting the necessity of investigating the plant responses to stress combinations27. Combined stresses can result in (even more) severe reductions in plant growth and determinantal effects on reproductive-related traits28. The impact of stress combination depends on the dominancy of each stress over the others, leading to enhanced or suppressed plant response (e.g., drought usually leads to stomata closure while stomata are open to allow cooling of leaf surface under heat stress). However, combined stress research is still emerging, and further investigations are required to understand better the complex regulation mediating plant responses under these conditions29. Thus, this study aims to highlight and recommend a phenotyping protocol using multiple imaging sensors that can be suitable to assess morpho-physiological responses and understand the underlying mechanisms of potato overall performance under single and combined stress treatments. As hypothesized, combining multiple imaging sensors proved to be a valuable tool to characterize the early and later strategies during plant stress response. Optimizing image-based phenotyping protocol will be an interactive tool for plant researchers and breeders to find traits of interest for abiotic stress tolerance.
Improved advanced high-resolution imaging tools and computer vision techniques have enabled the rapid development of plant phenotyping to obtain quantitative data from massive plant images in a reproducible manner39. This study aimed to adapt and optimize high throughput image-based methodology using an array of currently available imaging sensors to monitor the dynamic responses of plants under single and combined abiotic stresses. A few critical steps of the applied approach require adjustments, including applying stress and selecting a suitable imaging protocol for the measurements. Using multiple sensors for image acquisition allows the quantification of key phenotypic traits (such as plant growth, photosynthetic efficiency, stomatal regulations, leaf reflectance, etc.). In addition, improves the understanding of how potato plants respond to different abiotic stresses. This is a key prerequisite for accelerating plant breeding projects to develop climate-tolerant genotypes40. The morphological responses to the induced stress depend on the development stage. For example, inducing stress at the stolon or tuber initiation stage inhibits leaf and plant development and limits the number of stolons, thereby reducing the final yield41. However, under unfavorable conditions, plants utilize stress responses as an adaptive response to prevent and repair stress-induced cellular damage42. Plants have adaptive mechanisms to avoid and tolerate stress conditions depending on the severity level43.
To understand the mechanisms of plants, inducing the appropriate duration and intensity of stress and determining the plant responses to stress by using imaging sensors is considered one of the critical steps. When several stresses coincide, the intensity of one stress can overrule the effect of the others depending on the combination, intensity, and duration of the stresses. Thus, the stress effects can add up, or opposing responses can (partially) cancel each other, ultimately resulting in positive or negative effects on plants. The protocol selected in this study was based on previous experience to ensure that sufficient stress levels were applied. For instance, the application of the drought stress was adjusted to a moderate level as in a previous experiment, the response was not different from control treatments at an early stage of stress based on chlorophyll fluorescence imaging. This is due to the occurrence of photorespiration that acts as an alternative sink for electrons in the thylakoid membrane and a protective mechanism for the photosystem II44,45. Under the combined stress response, plant exposure to a mild primary stressor could enhance tolerance to a following stressor, which can have a beneficial or detrimental impact46. In this study, a stronger response was observed under combined stress compared to individual drought stress. By investigating other physiological responses, the results showed an increase in ΔT (deltaT) under drought as stomata close to avoid excess water loss. In contrast, the reverse response was observed under heat stress where ΔT was lower compared to control reflecting stomata opening to enhance leaf cooling in accordance with the findings in wheat under combined heat and drought stress20. During waterlogging, the increase of ΔT due to stomatal closure resulted from oxygen deficiency in the soil and disruption of root water homeostasis, thereby lowering the transpiration stream with an increase in the ABA, a key hormone in water stress responses47.
In plant stress studies, the duration of stress and subsequent recovery treatments is directly proportional to the stress intensity. For instance, moderate drought stress, such as maintaining soil moisture at 20% field capacity (FC), induces reversible phenotypic changes that typically recover after a single day of re-irrigation. In contrast, severe stress conditions like waterlogging result in extensive phenotypic damage, necessitating a longer recovery period. Although standardizing treatment durations is ideal, the inherent variability in stress intensities must be accounted for in experimental design.
The second critical step is to select an appropriate protocol and optimize the settings for each sensor. Chlorophyll fluorescence is a powerful tool in determining the performance of photosynthetic apparatus under stress48. Different chlorophyll fluorescence measuring protocols can be selected with either light or dark-adapted plants depending on the research question and the experimental design49. In this study, the selected protocol (short light response) enables the determination of various traits, including Fv'/Fm', φPSII, and qL, which indicate the photosynthesis performance under different conditions50. Previous studies showed that the used protocol in high-throughput phenotyping is effective in investigating the photosynthetic efficiency of plants under different applications of stress treatments and discriminating between healthy and stressed plants14,20. Based on the experimental design, it is very critical to consider the duration of the selected protocol when measuring in a high throughput system with a high plant population. Thus, the chlorophyll fluorescence measurement on light-adapted plants using a short-time protocol was selected to discriminate responses under different treatments. Genotype-environment interactions can influence many phenotypic traits, which is critical during measurement12. It is essential to consider that the duration of the measurement should be completed in a short time to minimize the diurnal effect on photosynthetic limitations51.
Thermal IR imaging was used to determine the canopy temperature and understand the stomatal regulation under different treatments52. It is worth mentioning that technological optimization was used where the heating wall was located on the opposite side of the camera, and the wall's temperature was dynamically controlled and programmable. Thus, adjusting the background heated wall with integrated environmental sensors is necessary to properly select plants from the background by increasing the contrast of the background temperature over the temperature of the imaged object.
Even though image analysis is automated, adjusting RGB thresholding indexes is still required to obtain a proper binary mask in RGB imaging to precisely select plants53. In addition, choosing multiple angles is important for appropriately estimating quantitative parameters, including digital biomass and growth rate. In this study, three angles (0°, 120°, and 240°) on the RGB side view were selected and averaged to calculate the plant volume and relative growth rate accurately.
Depending on the spectral range, many physiological traits can be investigated using hyperspectral imaging54. It is necessary to determine which of the reflectance indices provides the necessary information and shows the response of plants under different conditions14. It is highly demanded in screening for tolerant varieties and plant phenotyping to determine the correlation between the hyperspectral indices and other physiological traits55. In this study, plants under waterlogging treatment showed a pronounced response in the chlorophyll content and photosynthetic efficiency from the VNIR imaging. Moreover, different responses were observed in the water index calculated from SWIR imaging under heat treatments and waterlogging due to different stomatal regulations and water content in the leaves.
Thus, these findings highlight the utility of such an approach after optimizing the settings and the potential of using multiple sensors to find stress traits relevant to climate tolerance. Assessing the dynamics of the responses using multiple imaging sensors can be used as one of the powerful tools in improving breeding programs.
The authors have nothing to disclose.
This ADAPT project (Accelerated Development of multiple-stress tolerant Potato) has received funding from the European Union's Horizon 2020 research and innovation program under grant agreement No GA 2020 862-858. This work was partially supported by the Ministry of Education, Youth and Sports of the Czech Republic with the European Regional Development Fund-Project "SINGING PLANT" (no. CZ.02.1.01/0.0/0.0/16_026/0008446). The Core Facility Plants Sciences of CEITEC MU is acknowledged for its cultivation facility support. We acknowledge Meijer BV for providing the in-vitro cuttings used in this study. We thank Lenka Sochurkova for assisting in the graphical design of Figure 2 and Pavla Homolová for helping with the preparation of plant material during the experiments at Photon Systems Instruments (PSI) Research Center (Drásov, Czech Republic).
1.1” CMOS Sensor with RGB camera | PSI, Drásov, Czech Republic | https://psi.cz/ | The sensor delivers a resolution of 4112 × 4168 pixels for side view and 2560 × 1920 pixels for top view. The sensor is extremely sensitive and is a real megapixel CCD replacement and produces sharp, low-noise images |
FluorCam | PSI, Drásov, Czech Republic | FC1300/8080-15 | Pulse amplitude modulated (PAM) chlorophyll fluorometer |
Fluorcam 10 software | PSI, Drásov, Czech Republic | Version 1.0.0.18106 | For Chlorophyll fluorescence images visualization and analysis |
GigE PSI RGB – 12.36 Megapixels Camera | PSI, Drásov, Czech Republic | https://psi.cz/ | For the side view projections, line scan mode was used with a resolution of 4112 px/line, 200 lines per second. The imaged area from the side view was 1205 × 1005 mm (height × width), while the imaged area from the top view position was 800 × 800 mm. |
Hyperspectral Analyzer software | PSI, Drásov, Czech Republic | Version 1.0.0.14 | For hyperspectral images visualization and analysis |
Hyperspectral camera HC-900 Series | PSI, Drásov, Czech Republic | https://hyperspec.org/products/ | Visible-near-infrared (VNIR) camera 380-900 nm with a spectral resolution of 0.8 nm FWHM |
Hyperspectral camera SWIR1700 | PSI, Drásov, Czech Republic | https://hyperspec.org/products/ | Short-wavelength infrared camera (SWIR) camera 900 – 1700 nm with a spectral resolution of 2 nm FWHM |
InfraTec thermal camera (VarioCam HEAD 820(800)) | Flir, United States | https://www.infratec.eu/thermography/infrared-camera/variocam-hd-head-800/ | Resolution of 1024 × 768 pixels, thermal sensitivity of < 20 mK and thermal emissivity value set default to 0.95. with a scanning speed of 30 Hz and each line consisting of 768 pixels. The imaged area was 1205 × 1005 mm (height × width). |
LED panel | PSI, Drásov, Czech Republic | https://led-growing-lights.com/products/ | Equipped with 4 × 240 red-orange (618 nm), 120 cool-white LEDs (6500 K) and 240 far-red LEDs (735 nm) distributed equally over an imaging area of 80 × 80 cm |
Light, temperature and relative humidity sensors | PSI, Drásov, Czech Republic | https://psi.cz/ | Sensors used to monitor controlled conditions in greenhouse |
MEGASTOP Blue mats | Friedola | 75831 | To cover soil surface |
Morphoanalyzer software | PSI, Drásov, Czech Republic | Version 1.0.9.8 | For RGB images visualization and analysis and color segmentation analysis |
PlantScreen Data Analyzer software (Version 3.3.17.0) | PSI, Drásov, Czech Republic | https://plantphenotyping.com/products/plantscreen-modular-system/ | To visualize and analyze the data from all imaging sensors, watering-weighing unit and environmental conditions in greenhouse |
PlantScreen Modular system | PSI, Drásov, Czech Republic | https://plantphenotyping.com/products/plantscreen-modular-system/ | Type of phenotyping platform |
Plantscreen Scheduler software | PSI, Drásov, Czech Republic | Version 2.6.8368.25987 | To plan the experiment and set the measuring protocol |
SpectraPen MINI | PSI, Drásov, Czech Republic | https://handheld.psi.cz/products/spectrapen-mini/#details | Light meter to adjust light level on a canopy level |
TOMI-2 high-resolution camera | PSI, Drásov, Czech Republic | https://fluorcams.psi.cz/products/handy-fluorcam/ | Resolution of 1360 × 1024 pixels, frame rate 20 fps and 16-bit depth) with a 7-position filter wheel is mounted on a robotic arm positioned in the middle of the multi-color LED light panel with dimensions of 1326 x 1586 mm. |
Walk-in FytoScope growth chamber | PSI, Drásov, Czech Republic | https://growth-chambers.com/products/walk-in-fytoscope-fs-wi/ | Type of chambers used to grow the plant |
.