RootPainter: deep learning segmentation of biological images with corrective annotation.

Convolutional neural networks (CNNs) are a powerful tool for plant image analysis, but challenges remain in making them more accessible to researchers without a machine-learning background. We present RootPainter, an open-source graphical user interface based software tool for the rapid training of deep neural networks for use in biological image analysis. We evaluate RootPainter by training models for root length extraction from chicory (Cichorium intybus L.) roots in soil, biopore counting, and root nodule counting. We also compare dense annotations with corrective ones that are added during the training process based on the weaknesses of the current model. Five out of six times the models trained using RootPainter with corrective annotations created within 2 h produced measurements strongly correlating with manual measurements. Model accuracy had a significant correlation with annotation duration, indicating further improvements could be obtained with extended annotation. Our results show that a deep-learning model can be trained to a high accuracy for the three respective datasets of varying target objects, background, and image quality with < 2 h of annotation time. They indicate that, when using RootPainter, for many datasets it is possible to annotate, train, and complete data processing within 1 d.

Digging roots is easier with AI.

The scale of root quantification in research is often limited by the time required for sampling, measurement, and processing samples. Recent developments in convolutional neural networks (CNNs) have made faster and more accurate plant image analysis possible, which may significantly reduce the time required for root measurement, but challenges remain in making these methods accessible to researchers without an in-depth knowledge of machine learning. We analyzed root images acquired from three destructive root samplings using the RootPainter CNN software that features an interface for corrective annotation for easier use. Root scans with and without non-root debris were used to test if training a model (i.e. learning from labeled examples) can effectively exclude the debris by comparing the end results with measurements from clean images. Root images acquired from soil profile walls and the cross-section of soil cores were also used for training, and the derived measurements were compared with manual measurements. After 200 min of training on each dataset, significant relationships between manual measurements and RootPainter-derived data were noted for monolith (R2=0.99), profile wall (R2=0.76), and core-break (R2=0.57). The rooting density derived from images with debris was not significantly different from that derived from clean images after processing with RootPainter. Rooting density was also successfully calculated from both profile wall and soil core images, and in each case the gradient of root density with depth was not significantly different from manual counts. Differences in root-length density (RLD) between crops with contrasting root systems were captured using automatic segmentation at soil profiles with high RLD (1-5 cm cm-3) as well with low RLD (0.1-0.3 cm cm-3). Our results demonstrate that the proposed approach using CNN can lead to substantial reductions in root sample processing workloads, increasing the potential scale of future root investigations.

Simulating Root Growth as a Function of Soil Strength and Yield With a Field-Scale Crop Model Coupled With a 3D Architectural Root Model.

Accurate prediction of root growth and related resource uptake is crucial to accurately simulate crop growth especially under unfavorable environmental conditions. We coupled a 1D field-scale crop-soil model running in the SIMPLACE modeling framework with the 3D architectural root model CRootbox on a daily time step and implemented a stress function to simulate root elongation as a function of soil bulk density and matric potential. The model was tested with field data collected during two growing seasons of spring barley and winter wheat on Haplic Luvisol. In that experiment, mechanical strip-wise subsoil loosening (30-60 cm) (DL treatment) was tested, and effects on root and shoot growth at the melioration strip as well as in a control treatment were evaluated. At most soil depths, strip-wise deep loosening significantly enhanced observed root length densities (RLDs) of both crops as compared to the control. However, the enhanced root growth had a beneficial effect on crop productivity only in the very dry season in 2018 for spring barley where the observed grain yield at the strip was 18% higher as compared to the control. To understand the underlying processes that led to these yield effects, we simulated spring barley and winter wheat root and shoot growth using the described field data and the model. For comparison, we simulated the scenarios with the simpler 1D conceptual root model. The coupled model showed the ability to simulate the main effects of strip-wise subsoil loosening on root and shoot growth. It was able to simulate the adaptive plasticity of roots to local soil conditions (more and thinner roots in case of dry and loose soil). Additional scenario runs with varying weather conditions were simulated to evaluate the impact of deep loosening on yield under different conditions. The scenarios revealed that higher spring barley yields in DL than in the control occurred in about 50% of the growing seasons. This effect was more pronounced for spring barley than for winter wheat. Different virtual root phenotypes were tested to assess the potential of the coupled model to simulate the effect of varying root traits under different conditions.

Nutrient deficiency effects on root architecture and root-to-shoot ratio in arable crops.

Plant root traits play a crucial role in resource acquisition and crop performance when soil nutrient availability is low. However, the respective trait responses are complex, particularly at the field scale, and poorly understood due to difficulties in root phenotyping monitoring, inaccurate sampling, and environmental conditions. Here, we conducted a systematic review and meta-analysis of 50 field studies to identify the effects of nitrogen (N), phosphorous (P), or potassium (K) deficiencies on the root systems of common crops. Root length and biomass were generally reduced, while root length per shoot biomass was enhanced under N and P deficiency. Root length decreased by 9% under N deficiency and by 14% under P deficiency, while root biomass was reduced by 7% in N-deficient and by 25% in P-deficient soils. Root length per shoot biomass increased by 33% in N deficient and 51% in P deficient soils. The root-to-shoot ratio was often enhanced (44%) under N-poor conditions, but no consistent response of the root-to-shoot ratio to P-deficiency was found. Only a few K-deficiency studies suited our approach and, in those cases, no differences in morphological traits were reported. We encountered the following drawbacks when performing this analysis: limited number of root traits investigated at field scale, differences in the timing and severity of nutrient deficiencies, missing data (e.g., soil nutrient status and time of stress), and the impact of other conditions in the field. Nevertheless, our analysis indicates that, in general, nutrient deficiencies increased the root-length-to-shoot-biomass ratios of crops, with impacts decreasing in the order deficient P > deficient N > deficient K. Our review resolved inconsistencies that were often found in the individual field experiments, and led to a better understanding of the physiological mechanisms underlying root plasticity in fields with low nutrient availability.

Kinesthetic engagement in Gestalt evaluation outscores analytical 'atomic feature' evaluation in perceiving aging in crystallization images of agricultural products.

There is an increasing interest in a systemic approach to food quality. From this perspective, the copper chloride crystallization method is an interesting asset as it enables an estimation of a sample's 'resilience' in response to controlled degradation. In previous studies, we showed that an ISO-standardized visual evaluation panel could correctly rank crystallization images of diverse agricultural products according to their degree of induced degradation. In this paper we examined the role of contextual sensitivity herein, with the aim to further improve the visual evaluation. To this end, we compared subjects' performance in ranking tests, while primed according to two perceptional strategies (levels: analytical vs. kinesthetic engagement), according to a within-subject design. The ranking test consisted out of wheat and rocket lettuce crystallization images, exhibiting four levels of induced degradation. The perceptual strategy imbuing kinesthetic engagement improved the performance of the ranking test in both samples tested. To the best of our knowledge, this is the first report on the training and application of such a perceptual strategy in visual evaluation.

Bacteria utilizing plant-derived carbon in the rhizosphere of Triticum aestivum change in different depths of an arable soil.

Root exudates shape microbial communities at the plant-soil interface. Here we compared bacterial communities that utilize plant-derived carbon in the rhizosphere of wheat in different soil depths, including topsoil, as well as two subsoil layers up to 1 m depth. The experiment was performed in a greenhouse using soil monoliths with intact soil structure taken from an agricultural field. To identify bacteria utilizing plant-derived carbon, 13 C-CO2 labelling of plants was performed for two weeks at the EC50 stage, followed by isopycnic density gradient centrifugation of extracted DNA from the rhizosphere combined with 16S rRNA gene-based amplicon sequencing. Our findings suggest substantially different bacterial key players and interaction mechanisms between plants and bacteria utilizing plant-derived carbon in the rhizosphere of subsoils and topsoil. Among the three soil depths, clear differences were found in 13 C enrichment pattern across abundant operational taxonomic units (OTUs). Whereas, OTUs linked to Proteobacteria were enriched in 13 C mainly in the topsoil, in both subsoil layers OTUs related to Cohnella, Paenibacillus, Flavobacterium showed a clear 13 C signal, indicating an important, so far overseen role of Firmicutes and Bacteriodetes in the subsoil rhizosphere.