Measuring the Hidden Half of Forages
Roots are crucial for acquiring water and nutrients from the soil, preventing erosion, and building soil carbon (see “Why Roots Matter to Soil, Plants and You”). Furthermore, all these processes are also central to improving soil health (see “Look for These Soil Health Indicators in the Field”).
At the Noble Research Institute, we are addressing basic research questions about the genetics, development and physiology of roots while simultaneously applying this knowledge in forage breeding programs. We believe we can harness the power of roots to generate more nutrient-efficient, more resilient and more sustainable plant varieties that simultaneously increase producer profits while decreasing fertilizer inputs and pollution.
In root biology, the statement is cliché that roots are the hidden half. However, the description is appropriate. Because they are buried in soil, roots have been neglected in plant research due to a lack of technologies to efficiently study and understand roots. Therefore, a fundamental aspect of root research at Noble is to envision, invent and deploy novel methods for studying roots in the field.
Recently developed image-based methods that rely on root excavation are currently being used in the field, while research continues to develop futuristic technologies that will allow noninvasive scanning of roots in the field.
Elison Blancaflor, Ph.D., right, and Xiuwei Liu, Ph.D., demonstrate the “shovelomics” method for excavating crop root crowns.
The most commonly applied methods for studying roots in the field require removal of roots and soil, which are subsequently washed and photographed for further analysis. One of these commonly used methods has been called “shovelomics” and is appropriate for screening entire breeding populations of several hundred varieties in replicated plots. A researcher enters a research plot and uses a normal shovel to excavate the root crown, or the top portion of the root system. This root crown is soaked in water and gently washed with a water hose nozzle before being placed in a plastic bag and kept cold until photographing. Generally, this root crown breaks off the roots in the vertical direction but not in the horizontal as it is being removed from the soil. Though a lot of roots are certainly lost during excavation and washing, a substantial body of research now exists to show that traits measured from the root crown are useful for understanding the differences among varieties and why they perform differently.
The most advanced platform for imaging, or photographing, these root crowns was recently developed at Noble by the root phenomics laboratory, led by Larry York, Ph.D. The RhizoVision Crown platform integrates custom hardware and software in order to efficiently acquire root crown images that can easily be automatically processed using computer image analysis. The hardware platform consists of a backlight in front of which the washed root crown is suspended. A monochrome (black and white) machine vision camera is placed directly across from the roots. These cameras are normally used in factories to automate assembly lines, and they provide many useful features such as being rugged and easily programmable. An imaging software called RhizoVision Imager has been developed to allow a barcode reader to be used to scan the barcode on the sample, which triggers image acquisition and saves the file with the sample identification. The imager also allows live view and changeable camera settings. The software is versatile and is currently being used for multiple imaging platforms developed at Noble.
The RhizoVision Crown platform makes acquiring thousands of root images relatively easy. Due to the use of the backlight, these images contain a nearly black root crown on a nearly white background. Therefore, image analysis is greatly simplified because the computer can so easily identify the roots. Another software, RhizoVision Analyzer, was developed to efficiently process these thousands of images and extract a suite of nearly 30 features including root length, number of roots, root diameters and root angles, all of which are known to be important for vital root functions.
In collaboration with the mycology laboratory under the direction of Carolyn Young, Ph.D., the RhizoVision Crown platform was used to quantify the root system architectural differences between alfalfa plants that had survived cotton root rot and those that had not been affected. The results were consistent with survivors having lost their taproot but having the ability to compensate with many finer diameter lateral roots. Machine learning approaches were able to use the image-based data to predict whether roots were survivors or not with 75% accuracy. In another large collaboration involving small grains breeder Xuefeng Ma, Ph.D., at Noble and Felix Fritschi, Ph.D., soybean physiologist at University of Missouri, among others, the platform was shown to be able to identify wheat and soybean root crowns with 99% accuracy using machine learning. We have also identified substantial genetic variation for these root traits in several species and have found that many relate to plant performance, as measured with shoot mass. This platform is expected to have substantial impact in the plant sciences and breeding communities by making measuring the hidden half more accessible.
The destructive nature of currently used technologies means that we can’t follow the growth and development of roots in the field. We have to leave fields with holes from excavation and expend substantial manual labor in order to acquire data. To address these shortcomings, Noble Research Institute scientists are ramping up efforts to implement technology for nondestructive imaging of the complex world of roots underground. These nondestructive root-imaging initiatives involve techniques traditionally used in the biomedical or geophysical fields. For example, Noble researchers are collaborating with scientists at the Lawrence Berkeley National Laboratory (Berkeley Lab) and Subsurface Insights, LLC, a small business involved in software development for geophysical applications, on a project that uses electrical currents to image roots. The technology is analogous to a procedure called electroencephalography (EEG), which is used for imaging brains. In EEG, small metal discs with thin wires are attached to the scalp to detect brain electrical activity. Yuxin Wu, Ph.D., a geophysicist with the Climate and Ecosystem Sciences Division of the Berkeley Lab and Roelof Versteeg, Ph.D., founder of Subsurface Insights, LLC, coined the term Tomographic Electrical Rhizosphere Imaging (TERI) to refer to the EEG-like technology for measuring roots. The TERI technology involves injecting a small electrical current into the plant stem, which then travels throughout the root system. The TERI instrument is currently being refined so it can detect the electrical response of roots. The root electrical responses can then be translated into information about root mass, root surface area, rooting depth and root distribution in the soil. In addition to data on roots, TERI is expected to acquire data about soil texture and moisture content and to monitor how these soil variables change over time. The Advanced Research Projects Agency-Energy (ARPA-E) program of the Department of Energy funds Berkeley Lab, Subsurface Insights and Noble Research Institute scientists to develop TERI.
Ground penetrating radars (GPR) have also been tested at the Noble Research Institute to acquire information about roots in the soil without having to resort to digging up the plant. This geophysical technique uses electromagnetic radiation to locate objects beneath the ground. One major component of a GPR is an antenna that generates radar pulses, which are propagated through the ground. The radiation signals are reflected, absorbed or scattered by objects beneath the soil. A receiving antenna then detects variations to the radiation signals triggered by buried objects. Xiuwei Liu, Ph.D., a former Noble Research Institute postdoctoral fellow, who now leads his own group at the Chinese Academy of Sciences, collaborated with John Butnor from the U.S. Forest Service and Xuejun Dong, Ph.D., from Texas A&M AgriLife Research on applying GPR technology to detect roots of trees and wheat. Like the TERI technology described above, implementation of GPR on crops grown in agricultural fields will need additional years of refinement particularly with regard to correlating electrical and radiation signals with actual root distribution. A complicating factor in the implementation of these future technologies is the complexity of the soil. Nonetheless, these are the types of technologies that scientists at Noble and other research institutions worldwide are developing with the ultimate goal of breeding crops with improved root systems for more efficient capture of water and nutrients.
Rhizovision Crown Platform
The RhizoVision Crown platform represents a state-of-the-art instrument for high-throughput measures of crop root systems using hardware and software developed at Noble Research Institute.
Example data output from the RhizoVisoin Crown platform, including images with automated features overlaid on root images. A survivor alfalfa root crown (right) is compared to a nonaffected plant (left), and machine learning could identify these plants with substantial accuracy (middle).