- Open Access
High throughput phenotyping of morpho-anatomical stem properties using X-ray computed tomography in sorghum
© The Author(s) 2018
- Received: 2 August 2017
- Accepted: 4 July 2018
- Published: 13 July 2018
In bioenergy/forage sorghum, morpho-anatomical stem properties are major components affecting standability and juice yield. However, phenotyping these traits is low-throughput, and has been restricted by the lack of a high-throughput phenotyping platforms that can collect both morphological and anatomical stem properties. X-ray computed tomography (CT) offers a potential solution, but studies using this technology in plants have evaluated limited numbers of genotypes with limited throughput. Here we suggest that using a medical CT might overcome sample size limitations when higher resolution is not needed. Thus, the aim of this study was to develop a practical high-throughput phenotyping and image data processing pipeline that extracts stem morpho-anatomical traits faster, more efficiently and on a larger number of samples.
A medical CT was used to image morpho-anatomical stem properties in sorghum. The platform and image analysis pipeline revealed extensive phenotypic variation for important morpho-anatomical traits in well-characterized sorghum genotypes at suitable repeatability rates. CT estimates were highly predictive of morphological traits and moderately predictive of anatomical traits. The image analysis pipeline also identified genotypes with superior morpho-anatomical traits that were consistent with ground-truth based classification in previous studies. In addition, stem cross section intensity measured by the CT was highly correlated with stem dry-weight density, and can potentially serve as a high-throughput approach to measure stem density in grass stems.
The use of CT on a diverse set of sorghum genotypes with a defined platform and image analysis pipeline was effective at predicting traits such as stem length, diameter, and pithiness ratio at the internode level. High-throughput phenotyping of stem traits using CT appears to be useful and feasible for use in an applied breeding program.
- X-ray computed tomography
- High-throughput phenotyping
- Stem morphology
- Stem anatomy
- Stem biomechanics
- Computer vision
Breeding for standability and yield is a major focus of sorghum geneticists and breeders [1, 2]. Stem biomechanical and morpho-anatomical properties affect standability [3–7] and yield components in bioenergy sorghum  by influencing the plant’s ability to resist lodging and produce juicy and large stems. However, using existing assays to measure stem biomechanical and morpho-anatomical traits demands significant amounts of labor and time which reduce throughput. New high-throughput and advanced imaging technology provides a solution to alleviate this phenotyping bottleneck . This will ultimately enable plant scientists and breeders to evaluate larger segregating populations which would improve the selection process.
X-ray computed tomography (CT) has become a powerful tool for phenotyping plants and is becoming more widely available to a steadily growing number of plant biologists. As a result, this has led to vast amounts of image data which need to be efficiently managed, processed, mined, and analyzed [10–12]. Despite increasing interest in scanning plant stems using CT [13–15], there have been few studies to visualize and quantify in a high throughput manner above-ground structures of plants using CT.
Plant scientists have been using medical CT and industrial CT scanners to analyze a wide range of extant plant materials . Both scanners are based on the same underlying physics, but due to their difference in applications, industrial CT scanners offer a higher image resolution . Industrial CT scanners are often termed micro-CT or nano-CT because their resolution can range from 5 to 150 µm in the micro-CT and to as low as 0.5 µm in the nano-CT, compared to medical CT scanners, which have at best resolution of 70 µm . However, there are medical scanners available that can obtain similar resolution as industrial CT scanners. Regardless, the type of scanner being utilized, plant scientists are keen for scanners to be as high resolution as possible to accommodate small samples that require a high-resolution scan.
Given their resolution and capacity to detect external and internal phenotypic information in a non-invasive and non-destructive manner, combined with the ability to automate the process, has made the micro-CT the scanner of choice for plant studies . Micro-CT has been successfully used to characterize root structures, developing seeds, stems, leaves, and floral morphology and more at a very detailed level [14, 16–24]. However, depending on the resolution, the size of the sample, and desired signal-to-noise ratio, a CT scan may take several minutes to hours, and there is a sample size tradeoff [11, 17, 21]. Therefore, most studies to date, using micro-CT have been limited to greenhouse studies and used on small samples or small sample sizes that limit the throughput and applicability of this method in a large-scale field-breeding program. Nevertheless, these studies have provided numerous insights and methods to apply CT scanning and image numerous plants tissues.
In clinical research, a combination of biomechanics and X-ray CT has proven to be a powerful research technique to study whole-bone biomechanical properties [25, 26]. Application of such technology in crop improvement could be valuable as well. A study in maize successfully applied an X-ray CT to generate structural morphology of maize stems, which were then implemented in finite-element (FE) analyses. FE analyses performed to study the biomechanical response of these stems discovered that stem strength was highly dependent on stem morphology . The same group using dry maize stems grown under field conditions were able to scan up to 10 samples per run using X-ray CT, and identified a relationship between stem morphology and biomechanics in late-season stem lodging in maize .
In sorghum, stem lodging tends to occur at the grain filling stage  when there is significant moisture and turgor pressure that may affect biomechanical properties . As tissues mature and subsequently dehydrate as a result of senescence, the modulus of elasticity of these stem increases . Moreover, since bioenergy sorghum stem weight and moisture are good predictors of juice yield , it is important to evaluate plants when the physiological influences on the expression of these traits are minimal. Thus, previously mentioned results in late-season stem lodging in maize may not apply to bioenergy sorghum.
The current technological limitations of the micro-CT to acquiring plant morphological and anatomical data makes its application impractical in a field-breeding program. To be useful, the technology must have higher throughput when high-resolution is not needed. To address this problem, we propose the use of a medical CT scanner to visualize and quantify external and internal phenotypic structures in a high-throughput approach that would allow scanning larger samples and increase the number of samples per run of grass stems.
Using a medical CT has several advantages over a micro-CT when high-resolution is not necessary. For example, in a study by du Plessis et al. , the authors compared a medical CT to a micro-CT using samples of different densities. The authors concluded the medical CT scanners can produce useful data, significantly reduce scanning time, and provide an alternative for testing large numbers of samples when only moderate resolution is required. Medical CT can also scan larger samples than typical micro-CT systems that would be required to do in parts that would increase scanning time. Thus, using a medical CT would reduce the number of data sets, analysis, and computational power. In addition, industrial micro-CT scanners are not so easily accessible to many crop improvement programs and industrial micro-CT costs run much higher than medical CT systems .
Since the ultimate goal of plant biology is to map genotype to phenotype , high-throughput genotyping and phenotyping platforms must work in parallel with each other. A robust stem phenotyping platform would mitigate a phenotyping bottleneck existent in bioenergy/forage sorghums. The platform should accurately estimate stem geometry and morpho-anatomical traits, allow for a large number of samples to be run at the same time, fit large samples, produce acceptable repeatabilities for the traits, and work quickly with minimal effort. Thus, the objectives of this study were to (1) develop a practical high-throughput phenotyping platform and image data processing pipeline that can phenotype a large number of samples to extract stem morpho-anatomical properties and (2) validate the methodology.
Two separate field experiments were conducted in 2015 in College Station, Texas (30°33′05.6″N 96°26′14.8″W). Seeds were planted in one-row plots 5 m long and 0.76 m wide. Genotypes from Set 1 were arranged in a complete randomized block design. The target plant density was ~ 75,000 plants ha−1. For Set 2, F2 seeds were distributed in plots laid out in a row-by-column design. Seeds from Set 1 and Set 2 were sown in April. Agronomic practices standard for sorghum production in this area were used including irrigation as needed to minimize drought stress. Harvesting and evaluations occurred in July, approximately 95 days after planting.
For phenotyping each genotype in Set 1, six healthy plants were randomly selected from the middle of the plot and cut at the soil level. For Set 2, ten F2 plants were randomly selected from a ten plot population block. After harvest, any growth taller than 1.5 m was removed to fit the scanner and because stem lodging in sorghum occurs primarily between internodes three and six (which are typically between 0.5 and 1.5 m) . For most samples, the scanned section included internodes 1–7 and some genotypes had > 7 internodes in this section. This procedure was followed by the removal of leaf sheaf across the stem to get precise stem diameter measurements. During this time, samples were kept under moist conditions in a temperature-controlled environment at ~ 20 °C and then transported to be scanned. The process took under an hour after harvesting. After scanning the samples were stored again under the same conditions prior to measurements (maximum 6 h) to prevent tissue dehydration and to maintain natural material properties.
Phenotyping platform setup and CT-measurement
Visual stem pithiness measurements from Carvalho and Rooney  were used as these data included the same genotypes that were in Set 1. In brief, the percent of pithy stem cross-section area was visually estimated by using a rating scale system. This scale ranges from 1 to 9, where 1 corresponds to 90–100% pithiness and 9 to 0–10% pithiness. One unit increase in the scale equals to 10% decrease in the percent of the pithy area. For Set 2, the same protocol was followed, but in this case ratings were taken in the same plants that were scanned in the CT, and for internodes 3 and 6 only.
Computational image analysis
Next, the cross sections outlined previously were then used to determine the center, radius, diameter, rind area, intensity, and pithy area. CT intensity was measured as the ratio of the mean pixel intensities of a region and the maximum possible intensity of the image (which is 255). The rind area was defined as the area of the outer region. The inner circle was obtained by first excluding the pixels which intensities are higher than a threshold (175), and then fit the circle using the remaining pixel locations in that region. Finally, the percent pithy area was defined as the ratio between dark pixels (intensity is below a threshold, which is 20), inside each region and the area of the entire inner circle.
In total, over six morph-anatomical attributes were determined for each cross-section. Since it was not possible to detect nodes using the algorithm, node sections of the stem were added manually to the output. A separate function estimated internode length by multiplying the slice thickness of the CT image and the number of images within an internode section.
A total of ~ 500,000 images were produced by the CT scanning of ~ 150 plants. The preprocessing of phenotypic data involves removal of node sections since no data was collected at the nodes. Missing data was removed that may have occurred when the stems cross section move out of the area of estimation, or the algorithm did not detect a cross-section. Outlier detection was also performed by plotting the samples and identifying any extreme outliers.
The models were further evaluated by plotting predicted versus observed values in a 1:1 diagram of the model identified from the LOOCV method with the lowest RMSE. An R2 value close to 1.0 with a slope of observed versus predicted close to 1.0 and small RMSE values indicate that the model is precise with little bias . All statistical analyses were implemented in the R statistical language and computing environment .
Phenotypic variation for CT estimated traits was detected
Repeatability for CT estimates trait
Repeatabilities for CT-derived traits and ground-truth traits measured in 29 diverse sorghum genotypes
Second moment of an area (I)
Accuracy of estimating morph-anatomical traits using X-ray CT in sorghum
Correlations among CT-derived traits
Results demonstrate that CT estimates of morphological traits best correlated with biomechanical properties. Volume-CT was highly correlated with rigidity (r = 0.85; P < 0.001), respectively. Diameter-CT was positively correlated with rigidity (r = 0.71; P < 0.001) and negatively correlated with strength and stiffness (r = − 0.9; P < 0.001), (r = − 0.88; P < 0.001). These findings are consistent with results found by Gomez et al. .
Our image analysis pipeline could identify genotypes with superior morpho-anatomical traits that were consistent with ground-truth based classification previously performed by Carvalho and Rooney  and Gomez et al. . For example, the genotype Rio had smaller stem diameter than the genotype Tx13321 but longer internodes than Tx13321. Interestingly, the CT-estimated traits (internode diameter, internode length, and percent pithy area) were moderate to highly predictive of the manually collected traits. Percent pithy area explained 50% of the variation for pithiness rating. This value was much lower than the coefficient of determination found for the other traits evaluated. Stem pithiness occurs when the stem parenchyma cells die and are gradually filled with air creating a white, cottony, and pithy tissue. CT-derived percent pith is measured at every 0.6 mm slices across the length of the stem that can contribute to a larger variation than one visually collected pith rating. Therefore, one visual rating may not explain the variation across the length of the internode collected by the CT-based measurement and may have contributed to the lower R2 observed for stem pithiness in this study. Although the low variance explained for our pithiness rating may be a result of the visual score, visual ratings are one method to rate pithiness, and it has been demonstrated that using a flatbed scanner can increase the accuracy of phenotyping . Therefore, an association between CT-derived percent pith and percent pith area estimations using a flatbed scanner might result in a higher association.
In this study, CT estimated morphological traits had the strongest correlations with mechanical properties. This finding is consistent with a computation sensitivity analysis in maize by von Forell et al. , demonstrating that morphological traits demonstrate a stronger association with mechanical traits rather than tissue or material properties. The results from the computation sensitivity analysis were also consistent with a study using a micro-CT in maize . In our study, morphological measurements also had a strong effect on mechanical properties. For example, CT estimates of the second moment of an area were negatively correlated with stem strength and stiffness, demonstrating the strong effect stem morphology may have on mechanical traits. Similar results were reported by Gomez et al. . Furthermore, rind-CT was moderately associated with section modulus and is in line with a study by Ookawa et al.  where an indica variety of rice had strong culms due to a large section modulus that is associated with stem wall thickness. These results indicate that stem morphology has a strong effect on mechanical properties and morphological traits such as the second moment of an area and section modulus are to be considered when selecting for lodging resistance.
CT is based on the principle that the density of the tissue passed through by the X-ray beam can be measured by calculation of the attenuation coefficient . Therefore, material density is a major factor to consider when running plant samples in a CT scanner, as plant organs vary in tissue density. X-ray attenuation is mainly determined by the material properties of the plant tissues and can become visible by contrast according to density and atomic number of elements [12, 40]. Differences in X-ray attenuation in several plant stems were visibly apparent and primarily dependent on the anatomy, composition, and material density of the cross-section of the stem (i.e. rind is more lignified) (Fig. 3). At this attenuation level obtained by the SOMATOM Definition AS+ medical CT it is possible to detect the material density of the stems as well as rind and pithy area. It has been shown that medical CT scanners capture the changes in material density and composition of relative light and large objects [39, 41], such as stems of grasses. In this study, stem ‘density’ was estimated as the pixel intensities of a region and had a high correlation with internode dry weight density. Other studies indicate similar results . Intensity as used in this study, is a new method to quantify stem density in sorghum or other grass stems. Furthermore, in a recent study by the authors, it was found that internode density, volume, and stiffness can predict strength and can explain 75% of the variation . Therefore, using the methods developed in this study in combination with biomechanics can be used to apply selective breeding tools to improve lodging resistance.
The need for a high-throughput method for quantifying important morpho-anatomical traits related to stem lodging and juice yield in bioenergy sorghum motivated this pipeline and platform. While many high throughput methods are being developed to phenotype plants using unmanned aerial vehicle (UAV), robotics, and high throughput platforms such as the ARPA-E TERRA-REF project (http://terraref.org/) [43–46], none of these methods allow for combining external and internal stem phenotypic information of plants. Besides the potential applications discussed for our pipeline, it can also be applied to produce highly dimensional data used in 3D reconstruction and crop modeling (Additional file 2: Video S1).
Computer vision is an active and challenging field of computer science that is rapidly providing tools applicable to biological problems. In principle, images can be mined for phenotypes other than those which were collected . The spatial scanning resolution in an X-ray CT depends on the spot size of the X-ray source, the resolution of the X-ray detector, and used magnification of the system . While adding multiple samples in the medical CT may have introduced noise into the later measurements, we found that the image algorithm developed in this study was able to detect and extract useful external and internal phenotypic information effectively and accurately. However, the work herein is preliminary; there is room to improve on both processing and algorithms. For example, some of the coefficients of determination of the univariate regression did not explain all of the variation. We believe this was because plant stems vary in tissue density and the algorithm did not detect all cross-sections, therefore, it did not estimate all objects in the CT. Another reason may be that CT measurements were more precise and capture a larger portion of the variation than one single manually collected point that can also be subjective. We believe using computer vision and machine learning methods are warranted in future studies using CT in order to enable accurate phenotyping. In summary, the results indicate that medical CT scans can produce useful data in significantly reduced times, making it a good alternative for phenotyping plants.
The results herein indicate that CT-based estimates are associated with important traits in bioenergy/forage sorghum. Furthermore, predicting traits such as stem length, diameter, and pithiness ratio at the internode level by utilizing a high-throughput digital phenotyping approach using CT appears possible in an applied breeding program. Further work to improve algorithms and the accuracy of our models will enhance the speed and efficiency of this methodology allowing it to be applied to large populations, panels, and hybrids with high fidelity. As a selection tool, our protocol appears readily applicable in field-based and large-scale breeding programs.
FG and GC phenotyped the panel, performed data analysis, and were major contributors in writing the manuscript. FS and FG developed the computer algorithm used in this study to estimate CT traits. AM participated in the biomechanics study and editing of the manuscript. WL assisted in the design of the study and editing of the manuscript. All authors discussed the results and commented on the manuscript. All authors read and approved the final manuscript.
The authors are thankful to the staff of the Diagnostic Imaging & Cancer Treatment Center of the Texas A&M Veterinary Medicine & Biomedical Sciences facilities in College Station, Texas. FG is thankful to the Texas A&M University Louis Stokes Bridge to Doctorate Fellowship VII Award (No. 1249272) for a graduate fellowship and financial support granted to FG. GC is grateful to The National Council for Scientific and Technological Development (CNPq), Brazil, and The Tom Slick Graduate Research Fellowship at Texas A&M University for the Ph.D. fellowships and financial support provided.
The authors declare that they have no competing interests.
Availability of data and materials
The algorithm and datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.
Consent for publication
Ethics approval and consent to participate
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
- Mullet J, Morishige D, McCormick R, Truong S, Hilley J, McKinley B, Anderson R, Olson SN, Rooney W. Energy sorghum—a genetic model for the design of C4 grass bioenergy crops. J Exp Bot. 2014;65(13):3479–89. https://doi.org/10.1093/jxb/eru229.View ArticlePubMedGoogle Scholar
- Rooney WL, Blumenthal J, Bean B, Mullet JE. Designing sorghum as a dedicated bioenergy feedstock. Biofuels Bioprod Biorefin. 2007;1(2):147–57.View ArticleGoogle Scholar
- Niklas KJ. Plant biomechanics: an engineering approach to plant form and function. Chicago: University of Chicago press; 1992.Google Scholar
- Esechie HA, Maranville JW, Ross WM. Relationship of stalk morphology and chemical composition to lodging resistance in sorghum. Crop Sci. 1977;17(4):609–12.View ArticleGoogle Scholar
- Ookawa T, Hobo T, Yano M, Murata K, Ando T, Miura H, Asano K, Ochiai Y, Ikeda M, Nishitani R. New approach for rice improvement using a pleiotropic QTL gene for lodging resistance and yield. Nat Commun. 2010;1:132.View ArticlePubMedPubMed CentralGoogle Scholar
- Rutto LK, Xu Y, Brandt M, Ren S, Kering MK. Juice, ethanol, and grain yield potential of five sweet Sorghum (Sorghum bicolor [L.] Moench) cultivars. J Sustain Bioenergy Syst. 2013;03(02):113–8. https://doi.org/10.4236/jsbs.2013.32016.View ArticleGoogle Scholar
- Piñera-Chavez F, Berry P, Foulkes M, Molero G, Reynolds M. Avoiding lodging in irrigated spring wheat. II. Genetic variation of stem and root structural properties. Field Crops Res. 2016;196:64–74.View ArticleGoogle Scholar
- Carvalho G, Rooney WL. Assessment of stalk properties to predict juice yield in Sorghum. BioEnergy Res. 2017. https://doi.org/10.1007/s12155-017-9829-4.View ArticleGoogle Scholar
- Furbank RT, Tester M. Phenomics-technologies to relieve the phenotyping bottleneck. Trends Plant Sci. 2011;16(12):635–44.View ArticlePubMedGoogle Scholar
- Chen D, Neumann K, Friedel S, Kilian B, Chen M, Altmann T, Klukas C. Dissecting the phenotypic components of crop plant growth and drought responses based on high-throughput image analysis. Plant Cell. 2014;26(12):4636–55.View ArticlePubMedPubMed CentralGoogle Scholar
- Bucksch A, Atta-Boateng A, Azihou AF, Battogtokh D, Baumgartner A, Binder BM, Braybrook SA, Chang C, Coneva V, DeWitt TJ, Fletcher AG. Morphological plant modeling: unleashing geometric and topological potential within the plant sciences. Front Plant Sci. 2017;8:900.View ArticlePubMedPubMed CentralGoogle Scholar
- Metzner R, Eggert A, van Dusschoten D, Pflugfelder D, Gerth S, Schurr U, Uhlmann N, Jahnke S. Direct comparison of MRI and X-ray CT technologies for 3D imaging of root systems in soil: potential and challenges for root trait quantification. Plant Methods. 2015;11(1):17.View ArticlePubMedPubMed CentralGoogle Scholar
- Comparini D, Kihara T, Kawano T. Uses of X-ray 3D-computed-tomography to monitor the development of garlic shooting inside the intact cloves. Environ Control Biol. 2016;54(1):39–44.View ArticleGoogle Scholar
- Dhondt S, Vanhaeren H, Van Loo D, Cnudde V, Inzé D. Plant structure visualization by high-resolution X-ray computed tomography. Trends Plant Sci. 2010;15(8):419–22.View ArticlePubMedGoogle Scholar
- Robertson DJ, Julias M, Lee SY, Cook DD. Maize stalk lodging: morphological determinants of stalk strength. Crop Sci. 2017;57(2):926–34.View ArticleGoogle Scholar
- Hughes N, Askew K, Scotson CP, Williams K, Sauze C, Corke F, Doonan JH, Nibau C. Non-destructive, high-content analysis of wheat grain traits using X-ray micro computed tomography. Plant Methods. 2017;13(1):76.View ArticlePubMedPubMed CentralGoogle Scholar
- Cloetenes P, Mache R, Schlenker M, Lerbs-Mache S. Quantitative phase tomography of Arabidopsis seeds reveals intrecellular void network. Proc Natl Acad Sci. 2006;103:1426–14630.Google Scholar
- Kaminuma E, Yoshizumi T, Wada T, Matsui M, Toyoda T. Quantitative analysis of heterogenous spatial distribution of Arabidopsis leaf trichomes using micro X-ray computed tomography. Plant J. 2008;56:471–82.View ArticleGoogle Scholar
- Pajor R, Fleming A, Osborne CP, Rolfe SA, Sturrock CJ, Mooney SJ. Seeing space: visualization and quantification of plant leaf structure using X-ray micro-computed tomography: view point. J Exp Bot. 2013;64:385–90.View ArticlePubMedGoogle Scholar
- Rousseau D, Widiez T, Tommaso S, Rositi H, Adrien J, Maire E, Langer M, Olivier C, Peyrin F, Rogowsky P. Fast virtual histology using X-ray in-line phase tomography: application to the 3D anatomy of maize developing seeds. Plant Methods. 2015;11(1):55.View ArticlePubMedPubMed CentralGoogle Scholar
- Tracy SR, Gómez JF, Sturrock CJ, Wilson ZA, Ferguson AC. Non-destructive determination of floral staging in cereals using X-ray micro computed tomography (µCT). Plant Methods. 2017;13(1):9.View ArticlePubMedPubMed CentralGoogle Scholar
- Brereton NJB, Ahmed F, Sykes D, Ray MJ, Shield I, Karp A, Murphy RJ. X-ray micro-computed tomography in willow reveals tissue patterning of reaction wood and delay in programmed cell death. BMC Plant Biol. 2015;15(1):83.View ArticlePubMedPubMed CentralGoogle Scholar
- Dorca-Fornell C, Pajor R, Lehmeier C, Pérez-Bueno M, Bauch M, Sloan J, Osborne C, Rolfe S, Sturrock C, Mooney S. Increased leaf mesophyll porosity following transient retinoblastoma-related protein silencing is revealed by microcomputed tomography imaging and leads to a system-level physiological response to the altered cell division pattern. Plant J. 2013;76(6):914–29.View ArticlePubMedPubMed CentralGoogle Scholar
- Wang C-N, Hsu H-C, Wang C-C, Lee T-K, Kuo Y-F. Quantifying floral shape variation in 3D using microcomputed tomography: a case study of a hybrid line between actinomorphic and zygomorphic flowers. Front Plant Sci. 2015;6:724.PubMedPubMed CentralGoogle Scholar
- Keaveny TM. Biomechanical computed tomography—noninvasive bone strength analysis using clinical computed tomography scans. Ann N Y Acad Sci. 2010;1192(1):57–65.View ArticlePubMedGoogle Scholar
- Berger A. Bone mineral density scans. BMJ. 2002;325(7362):484.View ArticlePubMedPubMed CentralGoogle Scholar
- Von Forell G, Robertson D, Lee SY, Cook DD. Preventing lodging in bioenergy crops: a biomechanical analysis of maize stalks suggests a new approach. J Exp Bot. 2015;66(14):4367–71. https://doi.org/10.1093/jxb/erv108.View ArticleGoogle Scholar
- Gomez FE, Muliana AH, Niklas KJ, Rooney WL. Identifying morphological and mechanical traits associated with stem lodging in bioenergy Sorghum (Sorghum bicolor). BioEnergy Res. 2017. https://doi.org/10.1007/s12155-017-9826-7.View ArticleGoogle Scholar
- Niklas KJ, Spatz H-C. Plant physics. Chicago: University of Chicago Press; 2012.View ArticleGoogle Scholar
- du Plessis A, le Roux SG, Guelpa A. Comparison of medical and industrial X-ray computed tomography for non-destructive testing. Case Stud Nondestr Test Eval. 2016;6:17–25.View ArticleGoogle Scholar
- Hesse L, Wagner ST, Neinhuis C. Biomechanics and functional morphology of a climbing monocot. AoB Plants. 2016;8:plw005.View ArticlePubMedPubMed CentralGoogle Scholar
- Rowe NP, Isnard S, Gallenmüller F, Speck T. Diversity of mechanical architectures in climbing plants: an ecological perspective. Ecology and biomechanics: a mechanical approach to the ecology of animals and plants. Boca Raton: CRC Press; 2006. p. 35–59.Google Scholar
- Schulgasser K, Witztum A. On the strength of herbaceous vascular plant stems. Ann Bot. 1997;80(1):35–44.View ArticleGoogle Scholar
- Wagner ST, Isnard S, Rowe NP, Samain M-S, Neinhuis C, Wanke S. Escaping the lianoid habit: evolution of shrub-like growth forms in Aristolochia subgenus Isotrema (Aristolochiaceae). Am J Bot. 2012;99(10):1609–29.View ArticlePubMedGoogle Scholar
- Hallauer AR, Carena MJ, Miranda Filho J. Quantitative genetics in maize breeding, vol. 6. Berlin: Springer; 2010.Google Scholar
- Alam MM, van Oosterom EJ, Cruickshank AW, Jordan DR, Hammer GL. Predicting tillering of diverse sorghum germplasm across environments. Crop Sci. 2017;57(1):78–87.View ArticleGoogle Scholar
- Team RC. R: a language and environment for statistical computing. Vienna: R Foundation for Statistical Computing; 2014.Google Scholar
- Ookawa T, Aoba R, Yamamoto T, Ueda T, Takai T, Fukuoka S, Ando T, Adachi S, Matsuoka M, Ebitani T. Precise estimation of genomic regions controlling lodging resistance using a set of reciprocal chromosome segment substitution lines in rice. Sci Rep. 2016. https://doi.org/10.1038/srep30572.View ArticlePubMedPubMed CentralGoogle Scholar
- Lafond JA, Han L, Dutilleul P. Concepts and analyses in the CT scanning of root systems and leaf canopies: a timely summary. Front Plant Sci. 2015;6:1111.View ArticlePubMedPubMed CentralGoogle Scholar
- Plews AG, Atkinson A, McGrane S. Discriminating structural characteristics of starch extrudates through X-ray micro-tomography using a 3-D watershed algorithm. Int J Food Eng. 2009. https://doi.org/10.2202/1556-3758.1513.View ArticleGoogle Scholar
- Dutilleul P, Lontoc-Roy M, Prasher SO. Branching out with a CT scanner. Trends Plant Sci. 2005;9(411):412.Google Scholar
- Gomez FE, Muliana AH, Rooney WL. Predicting stem strength in diverse bioenergy sorghum genotypes. Crop Sci. 2018. https://doi.org/10.2135/cropsci2017.09.0588.View ArticleGoogle Scholar
- Watanabe K, Guo W, Arai K, Takanashi H, Kajiya-Kanegae H, Kobayashi M, Yano K, Tokunaga T, Fujiwara T, Tsutsumi N, Iwata H. High-throughput phenotyping of sorghum plant height using an unmanned aerial vehicle and its application to genomic prediction modeling. Front Plant Sci. 2017;8:421.View ArticlePubMedPubMed CentralGoogle Scholar
- Andrade-Sanchez P, Gore MA, Heun JT, Thorp KR, Carmo-Silva AE, French AN, Salvucci ME, White JW. Development and evaluation of a field-based high-throughput phenotyping platform. Funct Plant Biol. 2014;41(1):68–79.View ArticleGoogle Scholar
- Batz J, Méndez-Dorado MA, Thomasson JA. Imaging for high-throughput phenotyping in energy sorghum. J Imaging. 2016;2(1):4.View ArticleGoogle Scholar
- Barker J, Zhang N, Sharon J, Steeves R, Wang X, Wei Y, Poland J. Development of a field-based high-throughput mobile phenotyping platform. Comput Electron Agric. 2016;122:74–85.View ArticleGoogle Scholar
- Gehan MA, Kellogg EA. High-throughput phenotyping. Am J Bot. 2017;104(4):505–8.View ArticlePubMedGoogle Scholar