Biology subjects

Thorup-Kristensen, K.

Publications and source records attributed to Thorup-Kristensen, K..

9 recordsLinked to original sources

bioRxiv2026

Transformers Outperform ConvNets for Root Segmentation: A Systematic Comparison Across Nine Datasets

Root segmentation is a fundamental yet challenging task in image-based plant phenotyping. We present the first systematic comparison of Transformer and Convolutional Neural Network (ConvNet) architectures for root segmentation, evaluating 21 architectures across nine diverse datasets and comparing pre-trained models to training from scratch. Transformer-based models significantly outperform ConvNets for segmentation accuracy and root-diameter agreement. Pre-training significantly improves mean Dice from 0.623 to 0.666 (p = 3.3 x 10-10). We also find that Transformers benefit more from pre-training than ConvNets, with Dice improvements of +0.072 versus +0.022 (p = 3.7 x 10-4), supporting the hypothesis that fine-tuned Transformers transfer more effectively across large domain gaps. Among evaluated models, MobileSAM achieved the highest Dice score while maintaining computational efficiency. Dataset choice explained far more performance variance (70.9%) than model architecture (6.7%), suggesting that data curation matters more than model selection.

plant biology↗

bioRxiv2025

Five seasons with DeepRootLab: A unique facility for easier deep root research in the field

Deep-rooted crops accessing water and nutrients from deep soil layers enhance the resource base for crop production. However, studying these roots in field conditions is labor-intensive, limiting research scope. We established a field root research facility with 48 plots for replicated experiments. The facility includes 144 six-meter-long minirhizotron tubes and an AI-based pipeline for rapid root trait analysis. We also attempted to install access-tubes and customized ingrowth-core production for less-invasive root activity determination. Our study revealed significant differences in deep root density among species, particularly at depths of 2.5 to 4.5 meters, over five years. The less invasive studies using ingrowth-cores reached depths of 4.2 meters. Nutrient tracer 15N analysis showed marked differences in deep root activity among crop species. TDR sensors indicated varying water depletion in deeper soil layers, influenced by crop species and root growth patterns. We established a field facility for studying deep root growth and function, demonstrating its effectiveness in analyzing diverse deep-rooted plant species. This facility provides an ideal platform for conducting meaningful research in deep soil layers, yielding statistically and biologically significant results for agricultural applications.

plant biology↗

bioRxiv2022

Exploitation of neighbouring subsoil for nutrient acquisition under annual-perennial strip intercropping systems

Little is known of how the deep root systems of perennial crops contribute to deeper and better resource use when intercropped with annuals in arable fields. Therefore, we aimed at measuring the capacity of perennial deep roots, alfalfa (Medicago sativa L.) and curly dock (Rumex crispus L.) to access the nutrient source located under the neighboring annuals at 1.0 and 2.5 m of soil depth. Alfalfa and curly dock were able to access the tracer-labelled source placed at a distance under the annual crop strips. As a result, the reliance on deeper soil layer for nutrient uptake under intercroppings became greater compared with sole-croppings. Combination of an annual cereal (winter rye) and a perennial legume (alfalfa) with contrasting root systems exhibited higher resource complementarity compared with intercroppings having similar root systems or absence of legumes. Our results demonstrated that the deep-rooted perennials when intercropped with annuals can induce vertical niche complementarity, especially at deeper soil layers. This was assumed to be due to the vertically stratified root activity between the crop components, however, the magnitude of the effects depended on choice of crop combinations, and on types of tracers. Future studies should include estimates such as relative yield total and land equivalent ratio to quantitatively determine the effects of resource acquisition under annual-perennial intercropping in arable fields.

plant biology↗

bioRxiv2021

Dynamics of deep water and N uptake under varied N and water supply

AimsEnhanced nitrogen (N) and water uptake from deep soil layers may increase resource use efficiency whilst maintaining yield under stressed conditions. Winter oilseed rape (Brassica napus L.) can develop deep roots and access deep-stored resources such as N and water, while this potential has large uncertainties in variable environments. In this study, we aimed to evaluate the effects of reduced N and water supply on deep N and water uptake. MethodsOilseed rape plants grown in outdoor rhizotrons were supplied with 240 and 80 kg N ha-1 respectively in 2019 whereas a well-watered and a water-deficit treatment were established in 2020. To track deep water and N uptake, a mixture of 2H2O and Ca(15NO3)2 was injected into the soil column at 0.5 and 1.7 m depths. {delta}2H in transpiration water and {delta}15N in leaves were measured after injection. {delta}15N in biomass samples were also measured. ResultsDifferences in N or water supply had little effect on root growth. The low N treatment reduced water uptake throughout the soil profile, but caused a non-significant increment in 15N uptake efficiency at both 0.5 and 1.7 m. Water deficit in the upper soil layers led to compensatory deep water, while N uptake was not altered by soil water status. ConclusionOur findings demonstrate that for winter oilseed rape, high N application and water deficiency in shallow layers increases deep water uptake, and that the efficiency of deep N uptake is mainly sensitive to N supply rather than water supply.

plant biology↗

bioRxiv2020

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 (CNN) 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 minutes 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. 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.

plant biology↗

bioRxiv2020

RootPainter: Deep Learning Segmentation of Biological Images with Corrective Annotation

We present RootPainter, a GUI-based software tool for the rapid training of deep neural networks for use in biological image analysis. RootPainter facilitates both fully-automatic and semi-automatic image segmentation. We investigate the effectiveness of RootPainter using three plant image datasets, evaluating its potential for root length extraction from chicory roots in soil, biopore counting and root nodule counting from scanned roots. We also use RootPainter to compare dense annotations to corrective ones which are added during the training based on the weaknesses of the current model.

plant biology↗

bioRxiv2019

Optimizing root measurements in rhizotrons

Background and AimsThe line intersect method is widely used in rhizotron and minirhizotron studies to quantify roots and study cultivar and treatment differences in root growth. We investigated ways to optimize the line intersect method and root depth measurements with respect to data variability and the time spent on counting roots.\n\nMethodsRoot intensity was measured with three different grid patterns and different lengths of counting line on 2 m long transparent tube rhizotrons. Rooting depth was recorded by measuring the depth of the deepest root and by measuring the depth below which 5, 10 and 25 roots were observed.\n\nResultsFor root intensity the coefficient of variation (CV) was reduced 10-50 percentage points for grids that distributed counting lines equally across the measured area compared to using a restricted centralized area. In addition, the CV approached an asymptote of around 40 % when more than 50 root intersections per grid were observed. Further we show that recordings of the deepest root gave the most variance and least difference between means with a p-value of 0.65 for difference between cultivars. In contrast, a significant difference between cultivar rooting depths (p = 0.01) was found when using the depth below which 25 roots were observed.\n\nConclusionWe propose the use of grid designs adapted to different root densities to decrease time spent on counting roots at high root intensities, and minimize data variability at low root intensities. Further on rooting depth measurements including more roots may be a more useful parameter statistically to reveal variety or treatment differences in rooting depth.

plant biology↗

bioRxiv2019

The effect of drought and intercropping on chicory nutrient uptake from below 2 m studied in a multiple tracer setup

AimsWe tested if chicory acquires nutrients from soil layers down to 3.5 m depth and whether the deep nutrient uptake increases as a result of topsoil drought or topsoil resource competition. We also tested whether application of the trace elements Cs, Li, Rb, Sr, and Se, as tracers result in similar uptake rates. MethodsThe methodological tests were primarily carried out in a pilot experiment where the five tracers were applied to 1 m depth in lucerne and red beet grown in tube rhizotrons. The dynamics of deep nutrient uptake in chicory was studied in large 4 m deep rhizoboxes. A drought was imposed when roots had reached around 2 m depth. ResultsChicory acquired tracers applied to 3.5 m depth, but we found no compensatory tracer uptake with depth during drought. We found some indications of a compensatory tracer uptake from 2.3 and 2.9 m depth in intercropped chicory. Application of equimolar amounts of trace elements resulted in similar excess tracer concentrations within species. ConclusionChicory acquires nutrients from below 3 m but does not increase deep nutrient uptake as a response to limited topsoil nutrient availability.

plant biology↗

bioRxiv2018

Chicory demonstrates substantial water uptake from below 2 m, but still did not escape topsoil drought

AimsDeep-rooted agricultural crops can potentially utilize deep water pools and thus reduce periods where growth is water limited. Chicory (Cichorium intybus L.) is known to be deep-rooted, but the contribution of deep roots to water uptake under well-watered and drought conditions by the deep root system has not been studied. The aim of this study was to investigate whether chicory could reach 3 m depth within a growing season and demonstrate significant water uptake from the deeper part of the root zone. MethodsWe tested if chicory exposed to either topsoil drought or resource competition from the shallow-rooted species ryegrass (Lolium perenne L.) and black medic (Medicago lupulina L.) would increase deep water uptake in compensation for reduced topsoil water uptake. We grew chicory in 4 m deep soil filled rhizotrons and found that the roots reached 3 m depth within a growing season. ResultsWater uptake from below 1.7 m depth in 2016 and 2.3 m depth in 2017 contributed significantly to chicory water use. However, neither drought nor intercropping increased the deep water uptake. ConclusionChicory benefits from being deep-rooted during drought events, yet deep water uptake cannot compensate for the reduced topsoil water uptake during drought.

plant biology↗