Available via license: CC BY-NC-ND 4.0
Content may be subject to copyright.
Phytobiomes Journal •XXXX •XX:X-X https://doi.org/10.1094/PBIOMES-05-20-0042-RVW
REVIEW
Novel and Emerging Capabilities that Can Provide a Holistic
Understanding of the Plant Root Microbiome
Esther Singer,
1,2,†
John P. Vogel,
1,2,3
Trent Northen,
1,2
Christopher J. Mungall,
2
and Thomas E. Juenger
4
1
Joint Genome Institute, 1 Cyclotron Road, Berkeley, CA 94720
2
Lawrence Berkeley National Laboratory, 1 Cyclotron Road, Berkeley, CA 94720
3
Department of Plant and Microbial Biology, University of California, Berkeley, 111 Koshland Hall, Berkeley, CA 94720
4
The University of Texas at Austin, 2415 Speedway, Austin, TX 78712
Accepted for publication 3 November 2020.
ABSTRACT
In recent years, the root microbiome (i.e., microorganisms
growing inside, on, or in close proximity to plant roots) has been
shown to play an important role in plant health and productivity.
Despite its importance, the root microbiome is challenging to study
because of itscomplexity, heterogeneity, and subterranean location.
Fortunately, root microbiome research has seen a tremendous influx
of novel technologies (e.g., imaging tools, robotics, and molecular
analyses), experimental platforms (e.g., micro- and mesocosms),
and data integration, modeling, and prediction tools in the past
decade that have greatly increased our ability to dissect the complex
network of interactions between above- and belowground
environmental parameters, plants, bacteria, and fungi that dictate
soil and broader ecosystem health. Herein, we discuss methods that
are currently used in root microbiome research and that can be
expanded to phytobiome research in general ranging from
laboratory studies to mesocosm-scale studies and, finally, to field
studies; evaluate their relevance to ecosystem studies; and
discuss future root microbiome research directions.
Keywords: agriculture, ecology, metagenomics, microbiome,
mycology, nutrient cycling, plants, rhizosphere and phyllosphere,
soil ecology, soils
Root microbiota associate with every land plant and show
community compositions and dynamics that are distinct from the
surrounding soil microbial community (Lundberg et al. 2012). Both
rhizosphere (the immediate area around the roots) and root endo-
sphere (within the root) microbiomes affect plant health and soil
health via processes such as mineral and nutrient turnover (Alegria
Terrazas et al. 2016; Wei et al. 2019) and pathogen suppression
(Carri´
on et al. 2019). Attribution of specific processes to distinct
microbial players or populations is challenging because soil eco-
systems are among the most complex environments on Earth (Fierer
2017). Soils are made up of a multitude of heterogeneous abiotic
and biotic components that interact in a dynamic fashion over a
range of spatial and temporal scales (Fig. 1). Soils are categorized
based on their physical structure and their mineralogical compo-
sition. Soil type, together with climatic characteristics, allows for
the development and activity of biological constituents that are
specific to a given soil in a particular location and can vary dra-
matically among soils and locations (Fierer 2017). Those biological
constituents can include plants, insects, bacteria, archaea, and fungi,
which all contribute to and feed off of the biogeochemical cycles in
a given soil. The resulting complex network of interactions is
extremely challenging to disentangle due to technological limita-
tions and insufficient information in biological and chemical ref-
erence databases (Fierer 2017).
Furthermore, soils contain a vast diversity of microorganisms,
which are heterogeneously distributed and engage in frequent
horizontal gene transfer. Despite this, most root microbiome studies
present data from single time points or single locations and pri-
marily conduct amplicon sequencing combined with limited in-
formation on plant or environment. Although the average values
provided by such studies may suggest some interactions or
mechanisms, few studies follow up with the comprehensive sam-
pling necessary to definitively understand these mechanisms and
interactions. In addition, single-point studies are difficult to com-
pare or extrapolate to other environments or plants because
†
Corresponding author: E. Singer; esinger@lbl.gov
Funding: This work was supported by the U.S. Department of Energy (DOE) Joint
Genome Institute, a DOE Office of Science User Facility supported under contract
number DE-AC02-05CH11231, and was further supported by the Office of Science
(BER), U.S. DOE, Office of Science, Office of Biological and Environmental Re-
search, Genomic Science Program grant number DE-SC0014156 (to T. E. Juenger)
and the Trial Ecosystem Advancement for Microbiome Science and the Microbial
Community Analysis and Functional Evaluation in Soils Programs at Lawrence
Berkeley National Laboratory award DE-AC02-05CH11231 (to T. Northen); and
under the LBNL LDRD entitled “Improving biofuel crop yield from field to lab under
drought stress”(to E. Singer).
The author(s) declare no conflict of interest.
This article is in the public domain and not copyrightable. It may be freely reprinted with
customary crediting of the source. The American Phytopathological Society, 2021.
1
measured values can vary dramatically over time (e.g., some
bacteria bloom early in a growing season only to nearly disappear as
plants reach maturity) (Edwards et al. 2018).
Soil and other environmental characteristics can be important
indicators of biogeochemical processes that have occurred in the
past or are ongoing. Generally, few root and soil microbiome
studies take advantage of the relatively inexpensive techniques
(e.g., in situ soil sensors, determination of soil moisture via weight-
per-volume measurements, and pH strips) to measure soil char-
acteristics. Data on parameters such as pH, volumetric water
content, temperature, and salt concentration could allow researchers
to draw correlations between microbial activity, plant productivity,
and environmental parameters and facilitate opportunities to cross-
reference studies conducted under comparable conditions.
In the last decade, the root microbiome research community has
made tremendous progress in understanding the complexity of soil
ecosystems through improvements in experimental methods at both
laboratory and field scales. These exciting technological and sci-
entific advancements pave the way forward in root microbiome
research. This review summarizes recent technological advance-
ments and the resulting research opportunities categorized by
ecosystem component and scale (abiotic, biotic, space, and time),
and ends with an outlook and potential applications for phytobiome
research.
IMAGING TOOLS
Root structural imaging. Microbial colonization of the root and
rhizosphere can significantly affect root phenology and metabolism.
Roots demonstrate enormous phenotypic plasticity with respect to
anatomy, shape, cell type, cellular structure, metabolism, and
biochemical composition, and these characteristics contribute tre-
mendously to root exudation variation and, as a result, to microbial
community differentiation (Saleem et al. 2018). These reciprocal
interactions between roots and microbes are not well understood but
their direct link showcases the fact that, for understanding root
microbiomes, a foundational understanding of root biology is required.
Although hyperspectral imaging of leaves has been broadly
applied to monitor plant health, even simple imaging of intact roots
has lagged behind due to the challenges presented by the
opaqueness of soil (Bodner et al. 2017). Ideally, imaging of root
architecture, microbes, and chemical composition as well as vi-
sualization of fluxes such as carbon flow through plant compart-
ments and into the soil would be conducted at multiple temporal and
spatial scales. Most current methods for analyzing root growth
either require artificial growing conditions (e.g., hydroponics and
gels), are severely restricted in the fraction of roots detectable (e.g.,
rhizotrons), or are destructive (e.g., soil coring). For example,
many root phenotypic datasets have employed coring or “shov-
elomics”, subsequent root picking and washing, and imaging using
light imagers such as the RhizoVision Crown platform (Mattupalli
et al. 2019). This method provides valuable information about root
architecture; however, it is extremely laborious, it is often not
feasible to excavate deep roots, it can remain unknown how much
of the root system was recovered and scanned, and root excavation
often times terminates the experiment for the selected plants. All
of these methods are severely limited because they are destruc-
tive, low throughput, or artificial. The later point is particularly
important because root architecture can be significantly affected
by plant genetics, environmental conditions, soil type, and root-
colonizing bacteria and fungi (Gamalero et al. 2004; Mantelin et al.
2006).
Magnetic resonance imaging (MRI) presents a noninvasive
modality that addresses some of the limitations of other root
measurement techniques. When coupled with an analysis pipeline
in an automated system, MRI can monitor root mass, length, di-
ameter, tip number, growth angles (in two-dimensional polar co-
ordinates), and spatial distribution in a high-throughput manner
(van Dusschoten et al. 2016). Similarly, X-ray computed tomog-
raphy (CT) scanning can provide a comprehensive picture of root
systems as long as the roots have a diameter larger than the in-
struments’resolution (approximately 0.5 mm for medical CT
scanners) (Lin and Alessio 2009). Hence, small plants or young
roots are not likely to be resolved well. Another limitation common
to both MRI and CT technology is that plants must be grown in pots
that fit into the imaging machines and the applicability of MRI and
X-ray CT in three-dimensional (3D) imaging of root systems across
various pot sizes was recently evaluated (Metzner et al. 2015).
Although both MRI and CT were able to resolve high-quality 3D
images of root systems in vivo, the reconstructed length and image
details differed significantly between the two methods. In small
pots, CT outperformed MRI and provided more details thanks to
higher resolution whereas, in large pots, MRI was able to display
root systems more comprehensively than CT. Soil features such as
minerals and burrows can be resolved with CT, while MRI can
measure water content in roots and soil. Both CT and MRI,
struggled with roots thinner than 400 mm (Metzner et al. 2015).
Root vasculature imaging. There are a number of imaging tools
that can resolve root vasculature. Using Synchrotron X-ray
microtomography, Milien et al. (2012) contrasted the 3D images
of vascular systems of successful and unsuccessful graft interfaces
in vine rootstocks. Others have applied synchrotron X-ray
microtomography to visualize drought-induced embolism in vari-
ous plant species (Brodersen et al. 2013; Cochard et al. 2015;
Torres-Ruiz et al. 2015; Voltolini et al. 2017), to correlate root hair
with rhizosphere soil structure formation (Koebernick et al. 2017),
and to quantify root-induced changes of rhizosphere physical
properties (Aravena et al. 2013). Although synchrotron X-ray
micro-CT can render unprecedented detail into the microanat-
omy of plants and microorganisms, the focus window is relatively
limited (1 to 2 cm) and biological samples tend to lose viability as a
result of the intense X-ray radiation.
There are various other imaging methods that have been recently
developed or applied to phytobiome research, including super-
resolution confocal imaging, which can enhance 3D mapping of
root and microbial or fungal cells and showcase green fluorescent
proteins (Glaeser et al. 2016), and correlative confocal and focused
ion beam tool with integrated scanning electron microscope, which
allows for extremely fine-scaled 3D mapping (Lucas et al. 2014).
When applied individually or in combination, the abovementioned
imaging methods will provide opportunities to visualize plant tissue
and attached or internally residing bacteria, fungi, and viruses at
unprecedented resolution, as well provide information about their
physical and chemical context. Because root development is vital
for plant health, expansion of root image databases and novel
correlations between above- and belowground plant features will
enhance our understanding of plant response to environmental and
biological stimuli.
UNMANNED AERIAL VEHICLES
An important goal of the plant-microbiome field is to discover
beneficial or deleterious effects of microbes. This means that re-
cording and understanding plant phenotypes and linking them to
microbiome variation is key. Similarly, plant microbiomes are
intimately tied to the background soil; hence, monitoring soil
characteristics is important but can be challenging and labor-
intensive at appropriate temporal or spatial scales.
2
Unmanned aerial vehicles (UAVs) (e.g., “fixed wing”and
rotocopters) equipped with RGB cameras, infrared (IR) cameras,
multispectral and hyperspectral cameras, GPS, navigation systems,
programmable controllers, and automated flight planning have
emerged as powerful tools for nondestructive, high-throughput field
phenotyping that can be performed throughout the growth season
(Liebisch 2015). This has removed a bottleneck in phenotyping but
automated processing of this data still presents various challenges,
which are discussed elsewhere (Minervini et al. 2015). Monitoring
of agricultural fields using drones has become popular among
researchers (as well as agronomists, agricultural engineers, and
farmers) to more accurately plan and manage their experimental
operations. Drones can produce precise maps of soil characteristics
and plant characteristics (Bendig et al. 2014), as well as determine
irrigation needs, nitrogen levels, and pest occurrence (Huuskonen
and Oksanen 2018; Iost Filho et al. 2020). RGB, IR, and hyper- as
well as multispectral cameras attached to drones can collect images
of the aboveground portion in a range of wavelengths. The resulting
data can produce, for example, a vegetation index describing the
amount of wavelengths of light emitted from a crop and, hence, can
trigger irrigation systems or evaluate the sensitivity of crop breeds
to soil moisture in a high-throughput manner (Virlet et al. 2015).
Image data can also provide information about plant health status
over time and in dependence on the field location and, thereby,
allows the employment of an early warning and response system to
plant disease or stress (Mahlein 2016).
Traditionally, measuring soil quality parameters (e.g., nitrogen/
phosphorus/potassium ratios or organic matter content) requires
destructive sampling and laboratory analyses that are laborious, slow,
or expensive. Similarly, root phenotyping requires time- and labor-
intensive processing and scanning of root tissue to collect data such
as root length density and root architecture (as mentioned above).
Advances in imaging have been able to offset some of these hands-on
analyses: high-resolution RGB imaging can differentiate between soil
types facilitating soil type detection, which can improve mapping and
hence conservation efforts (Potter and Weigand 2018).
New approaches that overcome the limitations of laboratory tests
include thermal infrared imaging, which can be used to assess soil
moisture distribution and hydraulic properties (Boulet et al. 2009)
and inform land surface models (Kustas and Anderson 2009). Near-
infrared spectroscopy has been used for rapid and accurate iden-
tification of soil total nitrogen (TN), organic matter (OM), and pH
levels in soil that can replace laboratory techniques (He et al. 2007;
Ning et al. 2018). Similarly, hyperspectral imaging (HSI) can be
used to accurately provide TN, OM, and organic content infor-
mation in various soils (O’Rourke and Holden 2012) as well as
fungal viability based on pixel spectra specific to browned, dam-
aged, and undamaged tissue types (Taghizadeh et al. 2011). Be-
cause image processing of HSI is more challenging than that of
RGB imaging, the two technologies can be used in tandem; for
example, to optimize comprehensive analyses of soil and root
systems in rhizoboxes (Bodner et al. 2017). The accuracy of both IR
and HSI can be improved by applying extreme learning machine
models, which were previously used to increase the accuracy of soil
moisture and surface temperature measurements (Bai et al. 2015).
Because UAVs are scalable and programmable, we expect that
drone usage in phytobiome research will move toward autonomous
UAV fleets that can monitor extensive fields with an array of cheaper
and more accurate sensors. We also expect aerial monitoring to be
more closely coupled to robotics on the ground that could aid in
conducting soil and plant analysis and deployment and maintenance
of local sensor networks among various other tasks. Thus far, the
development of robotics to measure soil characteristics has primarily
focused on applications in environments that are difficult or unsafe to
access. For instance, a robot was developed for measuring soil
strength over depth, which is normally manually measured using a
penetrometer, in unsafe zones (Cao et al. 2003). The Mars Phoenix
Lander returned in situ measurements of Mars soil temperature,
generated a topography map using imaging, and excavated soil
samples for downstream testing (Arvidson et al. 2009).
PLANT AND MICROBIAL CHEMISTRY
Plant-microbiome signaling and metabolism rely on exchange of a
large diversity of metabolites derived from microorganisms, plants,
and the soil environment. Metabolomic methods enable direct
characterization of these small molecules from soils and the various
biological components. Given the large diversity of compounds that
reside intra- and extracellularly in these systems, mass spectrometry
(MS) coupled to chromatography such as liquid chromatography
(LC)-MS and gas chromatography (GC)-MS have become primary
methods for chemical analysis. Both techniques are well suited for
identification and quantification of a wide range of molecules found
in biological and environmental samples by coupling the physical
separation of the compounds using LC with the separation and
analysis of ions using MS mass (Jenkins et al. 2017). GC-MS and
LC-MS are complementary in many ways. GC-MS typically has
higher resolving power and produces richer fragmentation spectra,
which makes it particularly well suited for identifying molecules such
as small glycans that are often difficult to characterize by LC-MS. It is
also well suited for volatile molecules and poorly ionizing molecules
(e.g., hydrocarbons) that are often lost or not detected by LC-MS.
LC-MS, on the other hand, is better suited for thermally labile
compounds and is a technique of choice for analysis of novel
compounds. Typically, these approaches are suitable for identifica-
tion of several hundreds of metabolites based on spectral databases
and authentic standards (Vinaixa et al. 2016). However, they are
currently far from comprehensive, and improving metabolite iden-
tification is an important goal of metabolomics research.
A number of studies have used MS-based metabolomics to
examine the chemical exchanges within phytobiomes; for example,
the signaling molecules that direct the establishment of bacterial and
mycorrhizal pathogens or symbionts with host plants. A number of
metabolites have been identified, including sugars, amino acids,
organic acids, phenolic compounds, and plant hormones, that are
associated with beneficial interactions and are also implicated by
single-strain and whole-community approaches (O’Banion et al.
2020). Exometabolite profiling (characterization of the extracellular
metabolite pool) methods have been used to examine root exudates
and their function in recruiting soil bacteria (Zhalnina et al. 2018a).
O’Banion et al. (2020) have reviewed the function of the main
chemical constituents of plant microbe signaling. Similarly,
chemical imaging of solutes in soils has been reviewed (Santner
et al. 2015). Although MS imaging (untargeted investigation into
the spatial distribution of molecular species in a sample) is a
powerful and promising technique (Buchberger et al. 2018), it is
extremely difficult to identify organic components from complex
environmental samples due to chemical complexity of these
samples and the lack of physical separation of compounds prior to
ionization. New developments in using ion mobility to separate ions
within mass spectrometers have tremendous potential to overcome
these limitations and enable direct analysis of metabolites from
tissues and environmental samples (Spraggins et al. 2019).
PLANT GENETICS
It is well known that phytobiomes are affected by plant growth
form and life history (e.g., forbs, shrubs, and trees; annuals versus
3
perennials), plant community composition and habitat of origin,
and even host plant species (Fitzpatrick et al. 2018). In fact, there
is growing evidence of that intraspecific variability of plant hosts
produces variability in phytobiomes (Singer et al. 2018; Wagner
et al. 2016). Genetic differences within host species can affect
microbe recruitment, community assembly, and, ultimately, the
composition of phytobiomes. As such, the phytobiome can be
considered an extended phenotype of the plant that is determined
by host genetics, the environment, and their complex interaction.
Here, the standard tools of quantitative genetics can be used to
study the phytobiome. For example, family experimental designs
or kinship-based mixed models can be used to partition variation
in microbial abundance or composition into genetic and envi-
ronmental components of variance for an entire assemblage of
microbes associated with a particular plant compartment. This
approach can provide insight into the host genetic architecture of
the plant microbiome and, potentially, help to identify classes of
microbes with close affinities for specific genotypes within a
population.
A number of recent publications have documented genetic vari-
ation within plant species for aspects of the microbiome, including
providing estimates of heritability for overall microbial community
diversity and richness and for the abundance of specificmicrobial
taxa based oncounts derived from amplicon sequencing, for example
(Peiffer et al. 2013; Rodr´
ıguez et al. 2020; Walters et al. 2018) The
majority of such studies have focused on crop plants in agronomic
settings and little is known about the heritability of microbes from
more natural populations; one exception to this is the outdoor study of
Bergelson et al. (2019). We imagine that some of these host genetic
effects are related to available habitat for microbial establishment
(e.g., the extent of fine root growth), to resources shared with mi-
crobes as root exudates, or from more complex immune responses in
the plant. Incorporating host genetics in plant microbiome studies is
promising because it will point to mechanisms leading to beneficial or
deleterious plant–microbe interactions, as well as leverage the
growing resources available in plant genomics.
In order to more efficiently develop and deploy improved plant
varieties, it is valuable to identify the causal genes or genetic
markers underlying agronomic traits (e.g., yield) and disease re-
sistance (Visscher et al. 2017). In addition, there is a need to un-
derstand the plant genes that influence the composition and function
of the microbiome to improve our understanding and in order to
maximize plant productivity. Two methods are commonly used to
identify genes or markers associated with quantitative traits:
quantitative trait locus (QTL) mapping and genome-wide associ-
ation studies (GWAS). Both approaches rely on genome-wide scans
for statistical association between polymorphic genetic markers and
quantitative variation in a measured phenotype. In the case of
phytobiomes, the phenotype of interest could be a feature of the
aggregate microbial community (e.g., community richness and
evenness, principal coordinate analysis scores describing the mi-
crobial community) or an estimate of the relative abundance of a
specific taxon (Beilsmith et al. 2019). A key distinction between
these methods is that QTL mapping populations are derived from
lines crosses and, therefore, represent experimentally structured
populations, whereas GWAS focus on naturally occurring indi-
viduals. QTL mapping tends to have more power to detect true
associations but reduced ability to localize effects in the genome
because of limited recombination in a breeding population. In
contrast, GWAS are frequently underpowered, given limited
sample sizes, but can yield remarkably fine-scaled localization due
to extensive historical recombination. It can also be much faster to
establish a GWAS population than a QTL population because there
is no need to create recombinant progeny through complex breeding
designs across multiple generations. However, GWAS requires
dense markers and reliable controls for population structure and, at
best, yields correlative results rather than causal inference as
achieved with QTLs. Because, in QTL studies, fewer alleles and
markers are analyzed using a randomized genetic background,
statistical analysis can yield causal relationships between alleles and
traits (Lowry et al. 2015). Although both GWAS and QTL analyses
establishing relationships between plant genetics and phenotypic
traits are common, links between plant genetics and microbiome
composition and function have been rare. The earliest studies
utilizing this approach focused on plant-related microbial diseases
(Bartoli and Roux 2017), including fungal, oomycete, and bacterial
pathogens. More recently, studies utilizing the model plant
Arabidopsis thaliana have been published that explore complete
microbial communities based on 16S rRNA gene amplicon se-
quencing. For example, Horton et al. (2014) identified host loci that
influence fungal and bacterial colonization density on leaves across
an A. thaliana population in the field and found that loci encoding
defense and cell wall integrity affect bacterial and fungal com-
munity variation, whereas loci that influence the reproduction of
viruses, trichome branching, and morphogenesis affect bacterial
species richness. Similarly, Wallace et al. (2018) looked at the leaf
microbial communities across maize lines and found that functions
related to short-chain carbon metabolism, secretion, and nitro-
toluene degradation primarily encoded by Methylobacteria spp. are
heritable metabolic traits, and that few plant loci were found to be
significantly associated.
These studies provide an exciting glimpse of the potential im-
portance of host genetic variation in the phytobiome and give a clear
path to the identification of candidate genes. Future studies will help
to define the groups of microbes with strong host impacts, as well as
identify new genetic and metabolic pathways important in plant–
microbe interactions. Although aggregate community metrics may
be relatively straightforward to generate, they may be difficult to
interpret and less meaningful than studies focused on individual
microbial species. However, it is also unclear how to best define
microbial taxa for counting—what inference can be made from
amplicon sequence variants, traditionally defined operational tax-
onomic units, or gene content abundance derived from enrichment
or metagenomic analyses? Finally, genome-wide studies carry a
heavy multiple testing burden due to dense testing both across
genomes and also across multiple taxa or phenotypes. Care will
need to be taken to limit false positives and misleading infer-
ences—methods developed for other “omics”-based quantitative
genetic systems such as expression or metabolic QTL analyses may
provide helpful directions as the field matures.
FABRICATED ECOSYSTEMS
In an effort to conduct plant microbiome research across bio-
logically meaningful spatiotemporal scales and with increased
control, a range of fabricated ecosystems are being developed.
Experimental control and complexity are inversely related in plant
microbiome research. At the most extreme, controlled laboratory
experiments are often binary (one plant and one microbe), whereas
field experiments feature real-world complexity that is difficult to
replicate year by year. A new generation of experimental platforms
of increasing complexity now allows for multifactorial insight,
reproducibility, and increased statistical power.
The concept of controlled environments for exploring plant
ecophysiology dates back to the late 1940s, when Frits Went
developed a Phytotron at Caltech (Munns 2014), a “Climatron”in
St. Louis, MO (https://www.missouribotanicalgarden.org/gardens-
gardening/our-garden/gardens-conservatories/conservatories/
4
climatron.aspx), and an ecophysiology lab at the Desert Research
Institute, University of Nevada, Reno, which is now home to the
recently developed EcoCELLs (https://www.dri.edu/labs/ecocells/).
Went’s work inspired the development of the EcoTron program at
Centre National de la Recherche Scientifique, Montpellier, France
(https://www.ecotron.cnrs.fr/macrocosms/), and the EcoTron at
Imperial College London, United Kingdom (Fig. 2).
EcoTrons are large, fabricated ecosystems that consist of an
aboveground dome of approximately 40 m
3
and a belowground
chamber that contains a lysimeter that can hold 2 to 12 tons of soil
(up to 5-m
2
footprint and up to 2-m soil depth). The canopy area is
up to 2 m tall and allows work under natural light as well as under
controlled or artificial light conditions. Both above- and below-
ground compartments are equipped with arrays of sensors and
instrumentation for environmental control. Using the EcoTron,
simulations of a wide range of environmental scenarios under re-
alistic conditions can be performed, while measurements important
for ecosystem processes such as atmospheric and soil gas com-
position, temperature, and pH, among others, can be conducted.
Studies in EcoTrons will be increasing in the near future and will
provide unprecedented insights into ecosystem functioning; for
example, Roscher et al. (2019) found that the functional compo-
sition of communities is key in explaining carbon assimilation in
grasslands.
Mesocosms, which we call EcoPODs, are smaller versions of
EcoTrons with higher experimental throughput that bridge laboratory
and field studies (Eisenhauer and T¨
urke 2018). Existing EcoPODs
have a footprint of approximately 2.1 m
2
and can be filled with up to
1.23 m
3
of soil of 0.8 m in depth (UGT GmbH 2017). Using the
EcoPOD lysimeter technology, intact soil monoliths can be retrieved
from the field and studied under controlled conditions in the labo-
ratory. The aboveground portion is approximately 1.5 m tall and,
therefore, allows the study of a number of different plants in soil with
macro- and microorganisms in the context of environmental changes.
The contained nature of EcoPODs allows accurate mass balance
calculations (e.g., tracking of greenhouse gases over time). EcoPODs
allow precise conditioning of above- and belowground temperature
and moisture and, therefore, can simulate seasonal changes and
enable short- as well as long-term experiments. They are equipped
with state-of-the-art sensor technology allowing in situ measurements
of key environmental parameters, activities (fluxes) of organisms,
and ecosystems at micrometer to meter scales. EcoPODs can be
equipped with multi- and hyperspectral cameras that track plant
biomass and physiological states. In conjunction with highly con-
trolled physical and chemical conditions, researchers will be able to
track the microbial activity within the system using a variety of
genomic tools, including DNA or RNA shotgun metagenomics,
proteomics, and metabolomics. This will facilitate tracking of
Fig. 1. Overview of abiotic and biotic factors affecting plant, microbiome, and overall soil health.
5
microbial recruitment and activity across all life stages of the plant
and can simulate seasonal changes. Broadly, this system can be used
for fundamental research questions about biogeochemical cycles and
the role of biodiversity in ecosystem processes, as well as applied
studies that include biological or chemical components that require
increased safety clearance and cannot be easily tested in the field.
Because soil ecosystem and phytobiome experiments increasingly
rely on in situ sensing over time, EcoPODs can also serve as a testbed
for novel and improved sensing capabilities.
Complimenting approaches to develop more field-relevant
laboratory growth systems are composed of one or more sin-
gle plant chambers such as RootChips (Grossmann et al. 2011),
GLO-ROOT (Rell´
an- ´
Alvarez et al. 2015), EcoFAB (Zengler et al.
2019), and other systems that enable detailed characterization.
For example, RootChips systems provide a high-throughput
system for rhizosphere imaging, the GLO-ROOT systems en-
able direct imaging of root architecture within soils, and EcoFABs
are “fabricated ecosystems”that are aimed at creating model
Fig. 2. Most commonly used experimental platforms for plant microbiome studies displayed along a scale of experimental throughput versus relevance to
the field. Figure credit: EcoFAB: Dr. Lauren Jabusch; EcoTron: Dr. Sofie Thijs, Nathalie Beenaerts; and EcoPOD: Umwelt Ger¨ate Technik, Germany.
6
ecosystems on par with the model organisms used for genetic and
biological studies.
EcoFABs comprise a chamber, biological and abiotic compo-
nents (e.g., soil, plants, and microbes), and any measurement
technologies (e.g., sensors or microfluidic sampling apparatus).
EcoFABs allow real-time microscopy for high-resolution imaging
of plant root architecture and are currently designed to provide
sufficient materials for metabolomic, geochemical, and sequence-
based analyses. They are made using widely accessible 3D printing
technologies to fabricate controlled microbiome habitats that can be
standardized and easily disseminated between labs (Sasse et al.
2019; Zhalnina et al. 2018b). This approach provides flexibility that
enables scientists to add or change variables while monitoring
microorganisms and their interaction with plants. Replicability in
EcoFABs has recently been demonstrated in a ring trial across
multiple laboratories (Sasse et al. 2019). EcoFABs are also envi-
sioned to facilitate standardization of phytobiome research because
construction materials are cheap and construction instructions are
available (Gao et al. 2018).
Analogous to medical drug testing pipelines, which generally
begin as high-throughput laboratory screens and are gradually
scaled up to relevant mammal models and, finally, to human clinical
trials, we envision phytobiome research studies to similarly follow a
throughput versus relevance gradient from EcoFABs to EcoPODs
and, finally, to field studies (Fig. 2). Although this suite of fabri-
cated ecosystems is not aimed at simulating the real world, the
enhanced control over abiotic and biotic factors in these experi-
mental platforms enables plant root microbiome interaction studies
that are not possible in field experiments because fields generally
display greater complexity and unpredictability or do not allow for
manipulations. Thus, use of fabricated ecosystems can reveal im-
portant correlations and causations of individual metabolic reac-
tions as well as biogeochemical cycles. Challenges that have been
encountered or are foreseeable include the relatively short exper-
imental durations that can be executed in EcoFABs as well as
EcoPODs due to the size limitations of the respective platforms and
because of potential increases in parasite pressure as a result of air
and water flow limitations. On the other hand, EcoTrons are not set
up for quick turnover experiments and require expensive infra-
structure to start and end experiments. Although insights obtained
from greenhouse experiments have often not been replicable in the
field, we expect that EcoFAB can serve as a reproducible system, in
which microscopy and metabolomics can be applied to low-
complexity microbiomes in the context of plant roots. Data ob-
tained from individual microorganisms can inform microbially
based biogeochemical models, as discussed below. We expect
EcoPODs and EcoTrons to facilitate in situ sensing, climate ma-
nipulations, and deep soil monolith access. Links and extrapolations
among fabricated ecosystems and the field can be achieved by
generating and testing hypotheses across platform scales. For ex-
ample, field observations may be tested under replicable conditions
in EcoTron or EcoPOD and promising microbial candidates could
be isolated and further studied in EcoFABs. A reverse workflow is
also imaginable, where promising microbial isolates or plants
resulting from EcoFAB experiments may be tested in EcoPOD or
EcoTron before being potentially released into field experiments.
Furthermore, extrapolations could be testable beforehand by taking
advantage of archived datasets from sources such as long-term
observatories, including Neon (https://www.neonscience.org/).
Generally, challenges for extrapolating results of these fabricated
ecosystems to realistic field conditions could be presented by the
limited complexity in these laboratory systems; for example mi-
crobial isolates often perform predictably under laboratory con-
ditions but may be inactivated by night temperatures or competitors.
There is still a number of unknown unknowns which may sig-
nificantly affect plant performance, microbial community dynam-
ics, and soil nutrient cycling, and which vary from ecosystem to
ecosystem, hence resulting in a disconnect between studies con-
ducted in the laboratory versus in the field. Other challenges are
presented by natural climate variability in the field and the un-
certainty in climate change predictions, which are significantly
affected by socioeconomical drivers (Moss et al. 2010). Although
laboratory experiments may be conducted based on historic field
data or even in tandem with real-time field data measurements—for
example, using sensor platforms coupled to edge computing
(Beckman et al. 2016)—results may have limited applicability
under future climate scenarios. However, this is also true for re-
producibility of field experiments in general. Studying plant
microbiome interactions and soil processes under defined condi-
tions can assist in the identification and evaluation of such unknown
unknowns which, in turn, will improve applicability of laboratory
results to the field.
SYNTHETIC COMMUNITIES
Microbial communities found on healthy plants are incredibly
taxonomically diverse and include bacteria, archaea, fungi,
oomycetes, algae, protozoa, nematodes, and viruses (Bulgarelli
et al. 2013; Kemen 2014; Lu and Conrad 2005; Turner et al.
2013). This microbial complexity makes it impossible to defini-
tively establish causal relationships between plant and microbial
genotypes and phenotypes as well as environmental factors. In-
stead, representative synthetic communities (SynComs) of defined
complexity enable systematic bottom-up approaches in gnotobiotic
systems under controlled and reproducible conditions to determine
causal relationships (Lebeis et al. 2015).
In order to systematically test plant microbial community dy-
namics and functions in relation to the chemical composition of the
surrounding environment, comprehensive strain collections rep-
resenting the phylogenetic and functional diversity of the plant
microbiota have been established thanks to the cultivability of an
unexpectedly large fraction of the members of the plant microbiota
(M¨
uller et al. 2016). This high cultivability of plant-associated
bacteria is likely based on low-complexity food webs, continuous
substrate supply by the plant, and an essentially aerobic environ-
ment (M¨
uller et al. 2016). In addition to cultivation and subsequent
whole-genome analysis, screening SynComs of various complexity
for interactions and metabolic activity in correlation with envi-
ronmental parameters has been a bottleneck. Microfluidics tools
such as massively parallel on-chip coalescence of microemulsions
enable screening of 100,000 communities per day (Kehe et al.
2019). For example, bacterial isolates can be screened individually
and in combinatory sets as SynComs for various useful properties,
including plant-growth-promoting functions such as suppression of
pathogens or degradation of harmful substrates, for their potential in
biofuel production, or as environmental remediation agents. Such
tools coupled with high-throughput DNA or RNA sequencing and
long-read sequencing platforms including PacBio and Oxford
Nanopore (Sevim et al. 2019), as well as metabolomics and various
activity assays (Singer et al. 2019), now allow rapid profiling
composition, function, and activity of SynComs as well as complex
native microbial communities residing in soils and on plants.
DATA INTEGRATION, MODELING, AND PREDICTION
The quantity of data generated by the new technologies described
above surpasses the capabilities of traditional analysis methods.
Nevertheless, to gain insight, we need to integrate and fuse different
7
data streams. To accomplish this, we must overcome the hetero-
geneous data types and lack of standards for data exchange. Ul-
timately, we need systems that can dynamically pull in diverse data
from different devices and experimental modalities and intelligently
interpret it using background knowledge in order to derive new
hypotheses or make predictions such as being able to predict the
consequences of specific environmental changes on plant health
mediated by the microbiome.
Machine-learning (ML) methods and, in particular, deep learning
(DL) have proven particularly useful for classification problems
involving large datasets such as environmental data generated from
technologies, including thermal sensing and LiDAR. Supervised
ML techniques will learn to classify entities based on vectors of data
characteristics, trained from prelabeled data. DL techniques involve
the use of multilayer architecture neural networks (NNs). Different
DL architectures can be applied to different problems. Convolution
networks can be applied to image detection and recognition
problems (for example, detecting and segmenting the anatomical
components of a plant, and assessing its health), whereas recurrent
architectures such as long short-term memory can be applied to
time-series data. One of the challenges of phytobiome data are the
paucity of sample data or lack of resolution in imaging and in-
strumentation. One DL architecture designed to address this is the
generative adversarial network (GAN). A GAN can generate
plausible synthetic data by utilizing two NNs that are trained to-
gether in an adversarial scenario—one network (the discriminator)
attempts to distinguish real examples from fake ones, and the other
(the generator) creates plausible example data to fool the first. Over
time, both models improve, and the generated examples become
more plausible, reflecting real-world characteristics of the domain
without the need for explicit encoding of priors. In the context of
phytobiome data, GAN could, for instance, help to synthesize and
denoise imaging data (Yang et al. 2018). Although DL has seen
tremendous gains and achieved much over the last decade, there are
still a number of challenges. The input data must be in vector form,
which is straightforward for sensor data; however, complex bio-
logical information must be embedded in a suitable fashion. NNs
are famously inscrutable—they do not provide any explanation as to
why they produce a particular result. This is particularly prob-
lematic in the face of adversarial attacks, in which the NN is de-
liberately fooled by fake data designed to elicit a misclassification.
The burgeoning field of explainable artificial intelligence (Basu
et al. 2018) attempts to use a variety of techniques to make DL
decision making less of a black box process.
The field of DL and ML has seen a rapid advance in recent years
but, in many cases, DL methods may not yield improvements over
traditional methods. DL methods are best applied for complex
multidimensional data such as imaging data or for predictions
involving complex latent nonlinear mechanisms; for example, as
found in ecosystem models. Some have successfully applied DL
methods to modeling distinct ecosystem parameters such as soil
temperature over a soil depth profile (Gagne et al. 2020), and
processes such as ice-shelf melting as part of the Energy Exascale
Earth System Model (Caldwell et al. 2019). DL methods will also
gain importance in microbe-enabled soil biogeochemical models
that aim to predict links between climate change, elevated CO
2
concentrations, plant–microbe interactions, and soil nutrient cy-
cling (Georgiou et al. 2017; Tang and Riley 2015; Treseder et al.
2011; Zhu et al. 2016). For example, the ecosys model (https://
github.com/jinyun1tang/ECOSYS) allows for the incorporation of
microbially based models using traits such as growth rate (Vieira-
Silva and Rocha 2010), optimal temperature (Zeldovich et al.
2007), and resulting enzyme activity (Allison et al. 2010), as
well as genome size. Microbial traits, which can be obtained from
genomic data, help to identify and quantify trade-offs (Ferenci
2016). Trade-offs determine evolutionary adaptation constrains in
bacterial properties and prevent ecological fitness in all environ-
ments and, therefore, are important parameters to include in eco-
system models.
Most data have to be cleaned, filtered, integrated, and processed
before they can be utilized in ML or DL applications. This can be
particularly challenging for plant–microbe–soil ecosystem data,
which spans multiple scientific disciplines and technologies and
results in vastly different data types with large ranges of data
volume and temporal and spatial scales. The integration of these
data types and standardization of analysis has been notoriously
difficult, especially for data types that have been generated with
rapidly evolving technologies such as nucleic acid sequencing. This
difficulty in data integration combined with the incomplete col-
lection of datasets still poses challenges for the development and
application of current earth system modeling algorithms.
In order to compare between experiments, central and public data
storage is fundamental. The use of standardized metadata combined
with open controlled vocabularies or ontologies is crucial to being
able to interoperate between different data types. The “findable,
accessible, interoperable, reusable”(FAIR) data principles are
aimed at improving the data ecosystem to allow researchers to better
locate and integrate data. In the phytobiome sphere, the National
Microbiome Data Collaborative (Wood-Charlson et al. 2020) is a
new initiative to make microbiome data FAIR and aims to use
standards such as the Environment Ontology to describe envi-
ronmental characteristics of samples and the microbial ecosystems
embedded within them. One of those limitations surrounding ref-
erence databases is the paucity of experimentally validated data that
links microbial and plant metabolism, protein function, and DNA
sequence. As an example, whereas microbial genes are assigned
putative functions based on sequence homology, their actual ac-
tivity may deviate from these annotations leading, to incorrect
interpretations and predictions of ecosystem function (U.S. DOE
2019).
Furthermore, standardized analysis is critical and can be achieved
by using centrally updated, state-of-the-art software tools. KBase,
the U.S. Department of Energy Systems Biology Knowledgebase,
has offered a public data storage and analysis dashboard that allows
the generation of so-called narratives in which a dataset undergoes a
string of analyses (Arkin et al. 2018). In addition, large-scale field
datasets are increasingly taking advantage of supercomputer re-
sources and ML algorithms that are required to filter noise and
generate sensible interpretations from billions of data points.
OUTLOOK
The development of the abovementioned technologies and ex-
perimental platforms will improve our understanding of the
plant–microbe–atmosphere–soil ecosystem at high spatial and
temporal resolution. The combination and integrated use of the
discussed tools will further provide opportunities for novel ap-
proaches to plant root microbiome research. An example of an
integrated approach is the combination of UAVs equipped with
advanced imaging capabilities to study QTL or GWAS populations
growing in the field. This would streamline and scale current ex-
perimental procedures, so that new genetic markers for various
above- and belowground phenotypic characteristics could be
identified. These, in turn, could be correlated to microbiome
community profiles in roots and leaves.
Another example is the combined usage of SynComs and single
plants in EcoFabs for advanced root and microbe imaging resolved
over space and time complemented with metabolite analysis,
8
enabling systematic examination of the role of specific microbes
and metabolites in modifying root architecture. This approach can
help identify novel, specific microbial products that can be used to
influence important plant traits known to affect field performance
(e.g., root surface area). Microbial model systems can then be
engineered to produce promising compounds for tests on plants in
soil. Finally, we foresee EcoPODs and EcoTrons being used for
time-series experiments that span several weeks and months,
possibly years, in which high-throughput omics together with
continuous environmental sensor measurements can provide in-
depth yet broad-scale datasets that can be used for training artificial
intelligence algorithms related to biogeochemical cycling in rela-
tionship to climate.
Due to the many direct and indirect ties between local plant–
microbe–soil–ecosystem well-being and systems-wide ecological
health, technological improvements in phytobiome research are
directly translatable to improvements in climate change research.
The abovementioned advances in instrumentation and methodology
push precision agriculture and precision phytobiome research
forward and allow for improved and more sustainable crop pro-
ductivity under rapidly changing and increasingly extreme climatic
conditions. These advances will have impacts in food and energy
security and biosafety as well as environmental conservation and
bioremediation.
ACKNOWLEDGMENTS
We thank N. Bouskill for discussion on modeling and V. Shah for
assistance with Figure 1.
LITERATURE CITED
Alegria Terrazas, R., Giles, C., Paterson, E., Robertson-Albertyn, S., Cesco, S.,
Mimmo, T., Pii, Y., and Bulgarelli, D. 2016. Plant–microbiota interactions as
a driver of the mineral turnover in the rhizosphere. Pages 1-67 in: Advances in
Applied Microbiology, Vol. 95. S. Sariaslani and G. M. Gadd, eds. Academic
Press, Elsevier, Cambridge, MA, U.S.A.
Allison, S. D., Wallenstein, M. D., and Bradford, M. A. 2010. Soil-carbon
response to warming dependent on microbial physiology. Nat. Geosci. 3:
336-340.
Aravena, J. E., Berli, M., Menon, M., Ghezzehei, T. A., Mandava, A. K.,
Regentova, E. E., Pillai, N. S., Steude, J., Young, M. H., Nico, P. S., and
Tyler, S. W. 2013. Synchrotron X-Ray microtomography—New means to
quantify root induced changes of rhizosphere physical properties. Pages
39-67 in: Soil–Water–Root Processes: Advances in Tomography and
Imaging. S. H. Anderson and J. W. Hopmans, eds. SSSA Special Publication
61. Soil Science Society of America, Inc., Madison, WI, U.S.A.
Arkin, A. P., Cottingham, R. W., Henry, C. S., Harris, N. L., Stevens, R. L.,
Maslov, S., Dehal, P., Ware, D., Perez, F., Canon, S., Sneddon, M. W.,
Henderson, M. L., Riehl, W. J., Murphy-Olson, D., Chan, S. Y., Kamimura,
R. T., Kumari, S., Drake, M. M., Brettin, T. S., Glass, E. M., Chivian, D.,
Gunter, D., Weston, D. J., Allen, B. H., Baumohl, J., Best, A. A., Bowen, B.,
Brenner, S. E., Bun, C. C., Chandonia, J.-M., Chia, J.-M., Colasanti, R.,
Conrad, N., Davis, J. J., Davison, B. H., DeJongh, M., Devoid, S., Dietrich,
E., Dubchak, I., Edirisinghe, J. N., Fang, G., Faria, J. P., Frybarger, P. M.,
Gerlach, W., Gerstein, M., Greiner, A., Gurtowski, J., Haun, H. L., He, F.,
Jain, R., Joachimiak, M. P., Keegan, K. P., Kondo, S., Kumar, V., Land, M.
L., Meyer, F., Mills, M., Novichkov, P. S., Oh, T., Olsen, G. J., Olson, R.,
Parrello, B., Pasternak, S., Pearson, E., Poon, S. S., Price, G. A.,
Ramakrishnan, S., Ranjan, P., Ronald, P. C., Schatz, M. C., Seaver, S. M. D.,
Shukla, M., Sutormin, R. A., Syed, M. H., Thomason, J., Tintle, N. L., Wang,
D., Xia, F., Yoo, H., Yoo, S., and Yu, D. 2018. KBase: The United States
Department of Energy Systems Biology Knowledgebase. Nat. Biotechnol.
36:566-569.
Arvidson, R. E., Bonitz, R. G., Robinson, M. L., Carsten, J. L., Volpe, R. A.,
Trebi-Ollennu, A., Mellon, M. T., Chu, P. C., Davis, K. R., Wilson, J. J.,
Shaw, A. S., Greenberger, R. N., Siebach, K. L., Stein, T. C., Cull, S. C.,
Goetz, W., Morris, R. V., Ming, D. W., Keller, H. U., Lemmon, M. T.,
Sizemore, H. G., and Mehta, M. 2009. Results from the Mars Phoenix Lander
robotic arm experiment. J. Geophys. Res. Planets 114:E00E02.
Bai, Y., Wong, M. S., Shi, W.-Z., Wu, L.-X., and Qin, K. 2015. Advancing of
land surface temperature retrieval using extreme learning machine and spatio-
temporal adaptive data fusion algorithm. Remote Sens. 7:4424-4441.
Bartoli, C., and Roux, F. 2017. Genome-wide association studies in plant
pathosystems: Toward an ecological genomics approach. Front. Plant Sci. 8:763.
Basu, S., Kumbier, K., Brown, J. B., and Yu, B. 2018. Iterative random forests to
discover predictive and stable high-order interactions. Proc. Natl. Acad. Sci.
U.S.A. 115:1943-1948.
Beckman, P., Sankaran, R., Catlett, C., Ferrier, N., Jacob, R., and Papka, M.
2016. Pages 1-3 in: Waggle: An open sensor platform for edge computing.
Pages 1-3 in: 2016 IEEE SENSORS. IEEE, Orlando, FL, U.S.A.
Beilsmith, K., Thoen, M. P. M., Brachi, B., Gloss, A. D., Khan, M. H., and
Bergelson, J. 2019. Genome-wide association studies on the phyllosphere
microbiome: Embracing complexity in host-microbe interactions. Plant J. 97:
164-181.
Bendig, J., Bolten, A., Bennertz, S., Broscheit, J., Eichfuss, S., and Bareth, G.
2014. Estimating biomass of barley using crop surface models (CSMs)
derived from UAV-based RGB imaging. Remote Sens. 6:10395-10412.
Bergelson, J., Mittelstrass, J., and Horton, M. W. 2019. Characterizing both
bacteria and fungi improves understanding of the Arabidopsis root
microbiome. Sci. Rep. 9:24.
Bodner, G., Alsalem, M., Nakhforoosh, A., Arnold, T., and Leitner, D. 2017.
RGB and spectral root imaging for plant phenotyping and physiological
research: Experimental setup and imaging protocols. J. Vis. Exp. 126:e56251.
Boulet, G., Mougenot, B., and Abdelouahab, T. B. 2009. An evaporation test
based on thermal infra red remote-sensing to select appropriate soil hydraulic
properties. J. Hydrol. (Amsterdam, Neth.) 376:589-598.
Brodersen, C. R., McElrone, A. J., Choat, B., Lee, E. F., Shackel, K. A., and
Matthews, M. A. 2013. In vivo visualizations of drought-induced embolism
spread in Vitis vinifera. Plant Physiol. 161:1820-1829.
Buchberger, A. R., DeLaney, K., Johnson, J., and Li, L. 2018. Mass
spectrometry imaging: A review of emerging advancements and future
insights. Anal. Chem. 90:240-265.
Bulgarelli, D., Schlaeppi, K., Spaepen, S., van Themaat, E. V. L., and Schulze-
Lefert, P. 2013. Structure and functions of the bacterial microbiota of plants.
Annu. Rev. Plant Biol. 64:807-838.
Caldwell, P. M., Mametjanov, A., Tang, Q., Roekel, L. P. V., Golaz, J.-C., Lin,
W., Bader, D. C., Keen, N. D., Feng, Y., Jacob, R., Maltrud, M. E., Roberts,
A. F., Taylor, M. A., Veneziani, M., Wang, H., Wolfe, J. D., Balaguru, K.,
Cameron-Smith, P., Dong, L., Klein, S. A., Leung, R., Li, H.-Y., Li, Q., Liu,
X., Neale, R. B., Pinheiro, M., Qian, Y., Ullrich, P. A., Xie, S., Yang, Y.,
Zhang, Y., Zhang, K., and Zhou, T. 2019. The DOE E3SM coupled model
version 1: Description and results at high resolution. J. Adv. Model. Earth
Syst. 11:4095-4146.
Cao, P. M., Hall, E. L., and Zhang, E. 2003. Soil sampling sensor system on a
mobile robot. Pages 304-310 in: Proc. Vol. 5267. Intelligent Robots and
Computer Vision XXI: Algorithms, Techniques, and Active Vision.
Photonics Technologies for Robotics, Automation, and Manufacturing,
Providence, RI, U.S.A.
Carri´
on, V. J., Perez-Jaramillo, J., Cordovez, V., Tracanna, V., de Hollander, M.,
Ruiz-Buck, D., Mendes, L. W., van Ijcken, W. F. J., Gomez-Exposito, R.,
Elsayed, S. S., Mohanraju, P., Arifah, A., van der Oost, J., Paulson, J. N.,
Mendes, R., van Wezel, G., Medema, M. H., and Raaijmakers, J. M. 2019.
Pathogen-induced activation of disease-suppressive functions in the
endophytic root microbiome. Science 366:606-612.
Cochard, H., Delzon, S., and Badel, E. 2015. X-ray microtomography (micro-
CT): A reference technology for high-resolution quantification of xylem
embolism in trees. Plant Cell Environ. 38:201-206.
Edwards, J. A., Santos-Medell´
ın, C. M., Liechty, Z. S., Nguyen, B., Lurie, E.,
Eason, S., Phillips, G., and Sundaresan, V. 2018. Compositional shifts in root-
associated bacterial and archaeal microbiota track the plant life cycle in field-
grown rice. PLoS Biol. 16:e2003862.
Eisenhauer, N., and T¨
urke, M. 2018. From climate chambers to biodiversity
chambers. Front. Ecol. Environ. 16:136-137.
Ferenci, T. 2016. Trade-off mechanisms shaping the diversity of bacteria. Trends
Microbiol. 24:209-223.
Fierer, N. 2017. Embracing the unknown: Disentangling the complexities of the
soil microbiome. Nat. Rev. Microbiol. 15:579-590.
Fitzpatrick, C. R., Copeland, J., Wang, P. W., Guttman, D. S., Kotanen, P. M.,
and Johnson, M. T. J. 2018. Assembly and ecological function of the root
microbiome across angiosperm plant species. Proc. Natl. Acad. Sci. U.S.A.
115:E1157-E1165.
9
Gagne, D. J., Christensen, H. M., Subramanian, A. C., and Monahan, A. H.
2020. Machine learning for stochastic parameterization: Generative
adversarial networks in the Lorenz ’96 model. J. Adv. Model. Earth Syst. 12:
e2019MS001896.
Gamalero, E., Trotta, A., Massa, N., Copetta, A., Martinotti, M. G., and Berta, G.
2004. Impact of two fluorescent pseudomonads and an arbuscular
mycorrhizal fungus on tomato plant growth, root architecture and P
acquisition. Mycorrhiza 14:185-192.
Gao, J., Sasse, J., Lewald, K. M., Zhalnina, K., Cornmesser, L. T., Duncombe, T.
A., Yoshikunio, Y., Vogel, J. P., Firestone, M. K., and Northen, T. R. 2018.
Ecosystem fabrication (EcoFAB) protocols for the construction of laboratory
ecosystems designed to study plant-microbe interactions. J. Vis. Exp. 134:
e57170.
Georgiou, K., Abramoff, R. Z., Harte, J., Riley, W. J., and Torn, M. S. 2017.
Microbial community-level regulation explains soil carbon responses to long-
term litter manipulations. Nat. Commun. 8:1223.
Glaeser, S. P., Imani, J., Alabid, I., Guo, H., Kumar, N., K¨
ampfer, P., Hardt, M.,
Blom, J., Goesmann, A., Rothballer, M., Hartmann, A., and Kogel, K.-H.
2016. Non-pathogenic Rhizobium radiobacter F4 deploys plant beneficial
activity independent of its host Piriformospora indica. ISME J. 10:871-884.
Grossmann, G., Guo, W.-J., Ehrhardt, D. W., Frommer, W. B., Sit, R. V., Quake,
S. R., and Meier, M. 2011. The RootChip: An integrated microfluidic chip for
plant science. Plant Cell 23:4234-4240.
He, Y., Huang, M., Garc´
ıa, A., Hern´
andez, A., and Song, H. 2007. Prediction of
soil macronutrients content using near-infrared spectroscopy. Comput.
Electron. Agric. 58:144-153.
Horton, M. W., Bodenhausen, N., Beilsmith, K., Meng, D., Muegge, B. D.,
Subramanian, S., Vetter, M. M., Vilhj´
almsson, B. J., Nordborg, M., Gordon,
J. I., and Bergelson, J. 2014. Genome-wide association study of Arabidopsis
thaliana leaf microbial community. Nat. Commun. 5:5320.
Huuskonen, J., and Oksanen, T. 2018. Soil sampling with drones and augmented
reality in precision agriculture. Comput. Electron. Agric. 154:25-35.
Iost Filho, F. H., Heldens, W. B., Kong, Z., and de Lange, E. S. 2020. Drones:
Innovative technology for use in precision pest management. J. Econ.
Entomol. 113:1-25.
Jenkins, S., Swenson, T. L., Lau, R., Rocha, A. M., Aaring, A., Hazen, T. C.,
Chakraborty, R., and Northen, T. R. 2017. Construction of viable soil defined
media using quantitative metabolomics analysis of soil metabolites. Front.
Microbiol. 8:2618.
Kehe, J., Kulesa, A., Ortiz, A., Ackerman, C. M., Thakku, S. G., Sellers, D.,
Kuehn, S., Gore, J., Friedman, J., and Blainey, P. C. 2019. Massively parallel
screening of synthetic microbial communities. Proc. Natl. Acad. Sci. U.S.A.
116:12804-12809.
Kemen, E. 2014. Microbe–microbe interactions determine oomycete and fungal
host colonization. Curr. Opin. Plant Biol. 20:75-81.
Koebernick, N., Daly, K. R., Keyes, S. D., George, T. S., Brown, L. K., Raffan,
A., Cooper, L. J., Naveed, M., Bengough, A. G., Sinclair, I., Hallett, P. D., and
Roose, T. 2017. High-resolution synchrotron imaging shows that root hairs
influence rhizosphere soil structure formation. New Phytol. 216:124-135.
Kustas, W., and Anderson, M. 2009. Advances in thermal infrared remote
sensing for land surface modeling. Agric. For. Meteorol. 149:2071-2081.
Lebeis, S. L., Paredes, S. H., Lundberg, D. S., Breakfield, N., Gehring, J.,
McDonald, M., Malfatti, S., del Rio, T. G., Jones, C. D., Tringe, S. G., and
Dangl, J. L. 2015. Salicylic acid modulates colonization of the root
microbiome by specific bacterial taxa. Science 349:860-864.
Liebisch, F., Kirchgessner, N., Schneider, D., Walter, A., and Hund, A. 2015.
Remote, aerial phenotyping of maize traits with a mobile multi-sensor
approach. Plant Methods 11:9.
Lin, E., and Alessio, A. 2009. What are the basic concepts of temporal, contrast,
and spatial resolution in cardiac CT? J. Cardiovasc. Comput. Tomogr. 3:
403-408.
Lowry,D.B.,Taylor,S.H.,Bonnette,J.,Aspinwall,M.J.,Asmus,A.L.,Keitt,T.
H., Tobias, C M., and Juenger, T. E. 2015. QTLs for biomass and developmental
traits in switchgrass (Panicum virgatum). BioEnergy Res. 8:1856-1867.
Lu, Y., and Conrad, R. 2005. In situ stable isotope probing of methanogenic
archaea in the rice rhizosphere. Science 309:1088-1090.
Lucas, M. S., Guenthert, M., Gasser, P., Lucas, F., and Wepf, R. 2014.
Correlative 3D imaging: CLSM and FIB-SEM tomography using high-
pressure frozen, freeze-substituted biological samples. Pages 593-616 in:
Electron Microscopy. J. Kuo, ed. Methods in Molecular Biology (Methods
and Protocols), Vol. 1117. Humana Press, Totowa, NJ, U.S.A.
Lundberg, D. S., Lebeis, S. L., Paredes, S. H., Yourstone, S., Gehring, J.,
Malfatti, S., Tremblay, J., Engelbrektson, A., Kunin, V., del Rio, T. G., Edgar,
R. C., Eickhorst, T., Ley, R. E., Hugenholtz, P., Tringe, S. G., and Dangl, J. L.
2012. Defining the core Arabidopsis thaliana root microbiome. Nature 488:
86-90.
Mahlein, A.-K. 2016. Plant disease detection by imaging sensors—Parallels and
specific demands for precision agriculture and plant phenotyping. Plant Dis.
100:241-251.
Mantelin, S., Desbrosses, G., Larcher, M., Tranbarger, T. J., Cleyet-Marel, J.-C.,
and Touraine, B. 2006. Nitrate-dependent control of root architecture and N
nutrition are altered by a plant growth-promoting Phyllobacterium sp. Planta
223:591-603.
Mattupalli, C., Seethepalli, A., York, L. M., and Young, C. A. 2019. Digital
imaging to evaluate root system architectural changes associated with soil
biotic factors. Phytobiomes J. 3:102-111.
Metzner, R., Eggert, A., van Dusschoten, D., Pflugfelder, D., Gerth, S., Schurr,
U., Uhlmann, N., and Jahnke, S. 2015. 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 11:17.
Milien, M., Renault-Spilmont, A.-S., Cookson, S. J., Sarrazin, A., and Verdeil,
J.-L. 2012. Visualization of the 3D structure of the graft union of grapevine
using X-ray tomography. Sci. Hortic. (Amsterdam) 144:130-140.
Minervini, M., Scharr, H., and Tsaftaris, S. A. 2015. Image analysis: The new
bottleneck in plant phenotyping [Applications Corner]. IEEE Signal Process.
Mag. 32:126-131.
Moss, R. H., Edmonds, J. A., Hibbard, K. A., Manning, M. R., Rose, S. K., van
Vuuren, D. P., Carter, T. R., Emori, S., Kainuma, M., Kram, T., Meehl, G. A.,
Mitchell, J. F. B., Nakicenovic, N., Riahi, K., Smith, S. J., Stouffer, R. J.,
Thomson, A. M., Weyant, J. P., and Wilbanks, T. J. 2010. The next generation
of scenarios for climate change research and assessment. Nature 463:747-756.
M¨
uller, D. B., Vogel, C., Bai, Y., and Vorholt, J. A. 2016. The plant microbiota:
Systems-level insights and perspectives. Annu. Rev. Genet. 50:211-234.
Munns, D. P. D. 2014. “The awe in which biologists hold physicists”: Frits
Went’sfirst phytotron at Caltech, and an experimental definition of the
biological environment. Hist. Philos. Life Sci. 36:209-231.
Ning, J., Sheng, M., Yi, X., Wang, Y., Hou, Z., Zhang, Z., and Gu, X. 2018.
Rapid evaluation of soil fertility in tea plantation based on near-infrared
spectroscopy. Spectrosc. Lett. 51:463-471.
O’Banion, B. S., O’Neal, L., Alexandre, G., and Lebeis, S. L. 2020. Bridging the
gap between single-strain and community-level plant-microbe chemical
interactions. Mol. Plant-Microbe Interact. 33:124-134.
O’Rourke, S. M., and Holden, N. M. 2012. Determination of soil organic matter
and carbon fractions in forest top soils using spectral data acquired from
visible–near infrared hyperspectral images. Soil Sci. Soc. Am. J. 76:586-596.
Peiffer, J. A., Spor, A., Koren, O., Jin, Z., Tringe, S. G., Dangl, J. L., Buckler, E.
S., and Ley, R. E. 2013. Diversity and heritability of the maize rhizosphere
microbiome under field conditions. Proc. Natl. Acad. Sci. U.S.A. 110:
6548-6553.
Potter, C., and Weigand, J. 2018. Imaging analysis of biological soil crusts to
understand surface heating properties in the Mojave Desert of California.
Catena 170:1-9.
Rell´
an- ´
Alvarez, R., Lobet, G., Lindner, H., Pradier, P.-L., Sebastian, J., Yee, M.-
C., Geng, Y., Trontin, C., LaRue, T., Schrager-Lavelle, A., Haney, C. A.,
Nieu, R., Maloof, J., Vogel, J. P., and Dinneny, J. R. 2015. GLO-Roots: An
imaging platform enabling multidimensional characterization of soil-grown
root systems. eLife 4:e07597.
Rodr´
ıguez, C. E., Antonielli, L., Mitter, B., Trognitz, F., and Sessitsch, A. 2020.
Heritability and Functional Importance of the Setaria viridis bacterial seed
microbiome. Phytobiomes J. 4:40-52.
Roscher, C., Karlowsky, S., Milcu, A., Gessler, A., Bachmann, D., Jesch, A.,
Lange, M., Mellado-V´
azquez, P., Strecker, T., Landais, D., Ravel, O.,
Buchmann, N., Roy, J., and Gleixner, G. 2019. Functional composition has
stronger impact than species richness on carbon gain and allocation in
experimental grasslands. PLoS One 14:e0204715.
Saleem, M., Law, A. D., Sahib, M. R., Pervaiz, Z. H., and Zhang, Q. 2018.
Impact of root system architecture on rhizosphere and root microbiome.
Rhizosphere 6:47-51.
Santner, J., Larsen, M., Kreuzeder, A., and Glud, R. N. 2015. Two decades of
chemical imaging of solutes in sediments and soils—A review. Anal. Chim.
Acta 878:9-42.
Sasse, J., Kant, J., Cole, B. J., Klein, A. P., Arsova, B., Schlaepfer, P., Gao, J.,
Lewald, K., Zhalnina, K., Kosina, S., Bowen, B. P., Treen, D., Vogel, J.,
Visel, A., Watt, M., Dangl, J. L., and Northen, T. R. 2019. Multilab EcoFAB
study shows highly reproducible physiology and depletion of soil metabolites
by a model grass. New Phytol. 222:1149-1160.
Sevim, V., Lee, J., Egan, R., Clum, A., Hundley, H., Lee, J., Everroad, R. C.,
Detweiler, A. M., Bebout, B. M., Pett-Ridge, J., G¨
oker, M., Murray, A. E.,
10
Lindemann, S. R., Klenk, H.-P., O’Malley, R., Zane, M., Cheng, J.-F.,
Copeland, A., Daum, C., Singer, E., and Woyke, T. 2019. Shotgun
metagenome data of a defined mock community using Oxford Nanopore,
PacBio and Illumina technologies. Sci. Data 6:285.
Singer, E., Bonnette, J., Kenaley, S., Woyke, T., and Juenger, T. E. 2019. Plant
compartment and genetic variation drive microbiome composition in
switchgrass roots. Environ. Microbiol. Rep. 11:185-195.
Spraggins, J. M., Djambazova, K. V., Rivera, E. S., Migas, L. G., Neumann, E.
K., Fuetterer, A., Suetering, J., Goedecke, N., Ly, A., Van de Plas, R., and
Caprioli, R. M. 2019. High-performance molecular imaging with MALDI
trapped ion-mobility time-of-flight (timsTOF) mass spectrometry. Anal.
Chem. 91:14552-14560.
Taghizadeh, M., Gowen, A. A., and O’Donnell, C. P. 2011. The potential of
visible-near infrared hyperspectral imaging to discriminate between casing
soil, enzymatic browning and undamaged tissue on mushroom (Agaricus
bisporus) surfaces. Comput. Electron. Agric. 77:74-80.
Tang, J., and Riley, W. J. 2015. Weaker soil carbon–climate feedbacks resulting
from microbial and abiotic interactions. Nat. Clim. Change 5:56-60.
Torres-Ruiz, J. M., Jansen, S., Choat, B., McElrone, A. J., Cochard, H.,
Brodribb, T. J., Badel, E., Burlett, R., Bouche, P. S., Brodersen, C. R., Li, S.,
Morris, H., and Delzon, S. 2015. Direct X-Ray microtomography observation
confirms the induction of embolism upon xylem cutting under tension. Plant
Physiol. 167:40-43.
Treseder, K. K., Balser, T. C., Bradford, M. A., Brodie, E. L., Dubinsky, E. A.,
Eviner, V. T., Hofmockel, K. S., Lennon, J. T., Levine, U. Y., MacGregor, B.
J., Pett-Ridge, J., and Waldrop, M. P. 2011. Integrating microbial ecology into
ecosystem models: Challenges and priorities. Biogeochemistry 109:7-18.
Turner, T. R., Ramakrishnan, K., Walshaw, J., Heavens, D., Alston, M.,
Swarbreck, D., Osbourn, A., Grant, A., and Poole, P. S. 2013. Comparative
metatranscriptomics reveals kingdom level changes in the rhizosphere
microbiome of plants. ISME J. 7:2248-2258.
UGT GmbH. 2017. EcoUnit (Ecotrons). UGT Umwelt Ger¨
ate Technik. https://
www.ugt-online.de/en/products/lysimeter-technology/ecotrons/ecounit/
U.S. DOE. 2019. Breaking the Bottleneck of Genomes. Understanding Gene
Function Across Taxa, Workshop Report DOE/SC-0199. United States
Department of Energy, Office of Science, Bethesda, MD, U.S.A. https://
genomicscience.energy.gov/genefunction/
van Dusschoten, D., Metzner, R., Kochs, J., Postma, J. A., Pflugfelder, D., B¨
uhler,J.,
Schurr, U., and Jahnke, S. 2016. Quantitative 3D analysis of plant roots growing
in soil using magnetic resonance imaging. Plant Physiol. 170:1176-1188.
Vieira-Silva, S., and Rocha, E. P. C. 2010. The systemic imprint of growth and
its uses in ecological (meta)genomics. PLoS Genet. 6:e1000808.
Vinaixa, M., Schymanski, E. L., Neumann, S., Navarro, M., Salek,R. M., and Yanes,
O. 2016. Mass spectral databases for LC/MS- and GC/MS-based metabolomics:
State of the field and future prospects. TrAC Trends Anal. Chem. 78:23-35.
Virlet, N., Lebourgeois, V., Martinez, S., Labbe, S., Costes, E., and Regnard,
J. L. 2015. Phenotyping the response of an apple tree hybrid population to soil
water constraint under field conditions: New insights brought by high
resolution imaging. Acta Hortic. 1099:879-886.
Visscher, P. M., Wray, N. R., Zhang, Q., Sklar, P., McCarthy, M. I., Brown, M.
A., and Yang, J. 2017. 10 Years of GWAS discovery: Biology, function, and
translation. Am. J. Hum. Genet. 101:5-22.
Voltolini, M., Tas¸, N., Wang, S., Brodie, E. L., and Ajo-Franklin, J. B. 2017.
Quantitative characterization of soil micro-aggregates: New opportunities
from sub-micron resolution synchrotron X-ray microtomography. Geoderma
305:382-393.
Wagner, M. R., Lundberg, D. S., del Rio, T. G., Tringe, S. G., Dangl, J. L., and
Mitchell-Olds, T. 2016. Host genotype and age shape the leaf and root
microbiomes of a wild perennial plant. Nat. Commun. 7:12151.
Wallace, J. G., Kremling, K. A., Kovar, L. L., and Buckler, E. S. 2018.
Quantitative genetics of the maize leaf microbiome. Phytobiomes J. 2:
208-224.
Walters, W. A., Jin, Z., Youngblut, N., Wallace, J. G., Sutter, J., Zhang, W.,
Gonz´
alez-Peña, A., Peiffer, J., Koren, O., Shi, Q., Knight, R., del Rio, T. G.,
Tringe, S. G., Buckler, E. S., Dangl, J. L., and Ley, R. E. 2018. Large-scale
replicated field study of maize rhizosphere identifies heritable microbes. Proc.
Natl. Acad. Sci. U.S.A. 115:7368-7373.
Wei, X., Zhu, Z., Wei, L., Wu, J., and Ge, T. 2019. Biogeochemical cycles of key
elements in the paddy-rice rhizosphere: Microbial mechanisms and coupling
processes. Rhizosphere 10:100145.
Wood-Charlson, E. M., Anubhav, Auberry, D., Blanco, H., Borkum, M. I.,
Corilo, Y. E., Davenport, K. W., Deshpande, S., Devarakonda, R., Drake, M.,
Duncan, W. D., Flynn, M. C., Hays, D., Hu, B., Huntemann, M., Li, P.-E.,
Lipton, M., Lo, C.-C., Millard, D., Miller, K., Piehowski, P. D., Purvine, S.,
Reddy, T. B. K., Shakya, M., Sundaramurthi, J. C., Vangay, P., Wei, Y.,
Wilson, B. E., Canon, S., Chain, P. S. G., Fagnan, K., Martin, S., McCue, L.
A., Mungall, C. J., Mouncey, N. J., Maxon, M. E., and Eloe-Fadrosh, E. A.
2020. The National Microbiome Data Collaborative: Enabling microbiome
science. Nat. Rev. Microbiol. 18:313-314.
Yang, Q., Yan, P., Zhang, Y., Yu, H., Shi, Y., Mou, X., Kalra, M. K., Zhang, Y.,
Sun, L., and Wang, G. 2018. Low-dose CT image denoising using a
generative adversarial network with Wasserstein distance and perceptual loss.
IEEE Trans. Med. Imaging 37:1348-1357.
Zeldovich, K. B., Berezovsky, I. N., and Shakhnovich, E. I. 2007. Protein and
DNA sequence determinants of thermophilic adaptation. PLOS Comput. Biol.
3:e5.
Zengler, K., Hofmockel, K., Baliga, N. S., Behie, S. W., Bernstein, H. C.,
Brown, J. B., Dinneny, J. R., Floge, S. A., Forry, S. P., Hess, M., Jackson, S.
A., Jansson, C., Lindemann, S. R., Pett-Ridge, J., Maranas, C., Venturelli, O.
S., Wallenstein, M. D., Shank, E. A., and Northen, T. R. 2019. EcoFABs:
Advancing microbiome science through standardized fabricated ecosystems.
Nat. Methods 16:567-571.
Zhalnina, K., Louie, K., Hao, Z., Mansoori, N., Nunes da Rocha, U., Shi, S.,
Cho, H., Karaoz, U., Loqu´
e, D., Bowen, B. P., Firestone, M. K., Northen, T.
R., and Brodie, E. L. 2018a. Dynamic root exudate chemistry and microbial
substrate preferences drive patterns in rhizosphere microbial community
assembly. Nat. Microbiol. 3:470-480.
Zhalnina, K., Zengler, K., Newman, D., and Northen, T. R. 2018b. Need for
laboratory ecosystems to unravel the structures and functions of soil microbial
communities mediated by chemistry. mBio 9:e01175-18.
Zhu, Q., Riley, W. J., Tang, J., and Koven, C. D. 2016. Multiple soil nutrient
competition between plants, microbes, and mineral surfaces: Model
development, parameterization, and example applications in several tropical
forests. Biogeosciences 13:341-363.
11
