Content uploaded by Yujia Luo
Author content
All content in this area was uploaded by Yujia Luo on Sep 23, 2022
Content may be subject to copyright.
Soil Biology and Biochemistry 174 (2022) 108831
Available online 15 September 2022
0038-0717/© 2022 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
Bacterial and fungal co-occurrence patterns in agricultural soils amended
with compost and bokashi
Yujia Luo
a
,
b
,
*
, Juan Bautista Gonzalez Lopez
b
, H. Pieter J. van Veelen
b
, Valentina Sechi
b
,
Annemiek ter Heijne
a
, T. Martijn Bezemer
c
,
d
, Cees J.N. Buisman
a
,
b
a
Environmental Technology, Department of Agrotechnology and Food Sciences, Wageningen University, P.O. Box 17, 6700, AA, Wageningen, the Netherlands
b
Wetsus, European Centre of Excellence for Sustainable Water Technology, Oostergoweg 9, 8911, MA, Leeuwarden, the Netherlands
c
Above-Belowground Interactions Group, Institute of Biology, Leiden University, P.O. Box 9505, 2300, RA, Leiden, the Netherlands
d
Netherlands Institute of Ecology (NIOO-KNAW), Department of Terrestrial Ecology, Droevendaalsesteeg 10, 6708, PB, Wageningen, the Netherlands
ARTICLE INFO
Keywords:
Organic amendments
Co-occurrence network
Inter-kingdom correlations
Field experiment
ABSTRACT
The living soil harbors a signicant number and diversity of bacteria and fungi, which are essential in sustaining
soil ecosystem functions. Most studies focus on soil bacteria or fungi, ignoring potential interrelationships be-
tween kingdoms that coevolve and synergistically provide ecosystem functions. In a seven-year agricultural eld,
we explored the effects of organic amendments (OAs; i.e., compost and bokashi) on intra- and inter-kingdom co-
occurrence networks of soil bacterial and fungal communities. We observed that OAs changed the co-occurrence
patterns of bacteria and fungi. Distinct network modules were observed in the unamended and amended soils,
and the physicochemical properties of the soil could partially explain the formation of these modules. We also
found that compost and bokashi increased the proportion of positive correlations, and this could reduce
competition among microorganisms for soil resources. Our study reveals that soil management with OAs affects
relationships between bacterial and fungal subpopulations that physically co-exist, cooperate, and compete in
non-random structured networks. It highlights that ecosystem functions may depend on inter-kingdom in-
teractions shaped by different amendments and their applied dose.
1. Introduction
Soil ecosystem functioning is a fundamental component of produc-
tive agriculture, largely driven by a myriad of positive and negative,
direct and indirect interactions among soil organisms (Wagg et al.,
2019). These soil communities generally consist of diverse bacterial and
fungal subpopulations that coevolve and synergistically provide
ecosystem functions (Fuhrman, 2009; Bezemer et al., 2010; Haq et al.,
2014). Intra- and inter-kingdom interactions cannot be understood only
based on richness, diversity, and community composition (Wagg et al.,
2019), but aggregated organisms that physically co-exist, cooperate and
compete in non-random structures that can be conceptually projected
and investigated as co-occurrence networks (Barber´
an et al., 2012; Lu
et al., 2013; Ma et al., 2016). While biological activity in agricultural
soils facilitates the availability of nutrients for crops from organic
amendments (OAs), it remains unclear if common OAs alter the struc-
ture and function of soil microbial networks.
The soil microbial biomass is viewed as the “eye of the needle” in the
soil, through which organic matter must pass (Jenkinson, 1977; Sparl-
ing, 1985). In this sense, soil microbes are sensitive to the quality and
distribution of soil organic pools. Studies have shown that abiotic var-
iables induced by OAs control soil bacterial or fungal community
composition (P´
erez-Piqueres et al., 2006; Lu et al., 2015; Dumontet
et al., 2017), which subsequently shapes the relationships between mi-
crobial taxa (Prosser et al., 2007). While several studies have shown that
the addition of OAs can affect co-occurrence patterns of soil bacterial or
fungal communities, these studies did not interrogate interrelationships
between kingdoms (Ling et al., 2016; Schmid et al., 2018; Xue et al.,
2018; Yang et al., 2019, 2020; Zhou et al., 2019).
Nonetheless, these studies revealed that the increased complexity of
the networks contributed to the stability and resilience of the native soil
microbiota (Yang et al., 2019). In addition, modules (a group of nodes
more densely connected to each other than to nodes outside the group)
in bacterial networks correlated differentially with specic soil variables
after applying different OAs (Ling et al., 2016). Here, a node refers to a
bacterial or fungal OTU. The connections of nodes in the network are
* Corresponding author. Environmental Technology, Wageningen University, P.O. Box 17, 6700, AA, Wageningen, the Netherlands.
E-mail addresses: yujia.luo@wetsus.nl, yujia.luo@outlook.com (Y. Luo).
Contents lists available at ScienceDirect
Soil Biology and Biochemistry
journal homepage: www.elsevier.com/locate/soilbio
https://doi.org/10.1016/j.soilbio.2022.108831
Received 27 August 2021; Received in revised form 5 September 2022; Accepted 9 September 2022
Soil Biology and Biochemistry 174 (2022) 108831
2
called edges, presenting positive or negative correlations between mi-
crobes. Exploring whether and how co-occurrences among bacteria and
fungi in the soil are shaped by commonly applied OAs should provide a
more integrated understanding of the soil microbiome and may provide
insights into ecosystem functioning for optimizing agricultural soil
management.
In this study, we report a seven-year eld experiment in which we
investigated how two OAs, compost, and bokashi, applied at high and
low doses, affected long-term soil bacterial and fungal intra- and inter-
kingdom relationships using co-occurrence network analysis. Compost
and bokashi are obtained from aerobic composting and fermentation,
respectively. Moreover, bokashi is prepared with the addition of com-
mercial microbial inoculum, which contains ve main groups of mi-
croorganisms, including lactic acid bacteria, photosynthetic bacteria,
yeasts, actinobacteria, and fermenting fungi (Shin et al., 2017). After a
seven-year annual application of compost and bokashi, we expect that
soil microorganisms co-occur in a non-random way rather than by
chance in the compost and bokashi amended soils, with a higher level of
non-random networks than the unamended soil. In addition, compost
and bokashi are very different regarding the stability of organic sub-
strates and nutrient status. The composting process leads to the removal
of labile organic fractions and the concentration of more recalcitrant
fractions, stabilizing organic matter with low nutrient content and
slow-release fertility (Neher et al., 2015). On the other hand, the C
content in bokashi can be more readily metabolized than that in compost
due to less efcient mineralization during fermentation than in com-
posting. Therefore, we expect that adding compost and bokashi will
shape the physicochemical properties of soil, and this will result in
distinct microbial modules. We also expect that adding compost and
bokashi will affect the intra- and inter-kingdom associations since
compost and bokashi provide organic substrates and nutrients. Antag-
onistic and mutualistic relationships have been reported between spe-
cic bacteria and fungi but depend on the local conditions (Nazir et al.,
2012; Fritsche et al., 2014). For instance, fungi can serve as potential
providers of nutrients and energy for bacteria in carbon- and
nutrient-limited soils (Haq et al., 2014). Therefore, we expect that in
soils amended with compost and bokashi, the inter-kingdom associa-
tions will be reduced compared with the unamended soil and that a
higher OA application dose will lead to a lower percentage of
inter-kingdom relations.
2. Materials and methods
2.1. Field experiment description
The eld experiment was initiated in autumn 2013 by Agriton
(Noordwolde, The Netherlands) at the experimental farm Ebelsheerd of
Stichting Proefboerderijen Noordelijke Akkerbouw (SPNA, Nieuw
Beerta, The Netherlands). The farm eld soil consists of 46% clay, 37%
silt, and 12% sand with a pH of 8.2. The soil contains 8.8% organic
matter. The area is characterized as strong swelling and shrinking soils
with high soil compaction (Bronswijk and Evers-Vermeer, 1990; Roelf-
sema, 2010). Compost and bokashi were applied yearly before winter
wheat (Triticum aestivum), with low and high application doses
(Table 1). Compost and bokashi were prepared from the same amount of
grass that underwent composting and anaerobic fermentation, respec-
tively. The commercial microbial inoculum (EM-1
®
, Agriton, The
Netherlands) was used to start the fermentation process when preparing
Bokashi. All soils received two levels of articial N fertilization. Table 1
shows the arrangement of the eld experiment.
Soil cores (0–15 cm deep, 3 cm diameter) were collected from each
plot in October 2020. Soil samples were transported to the laboratory on
ice, and each soil sample was divided into three subsamples afterward.
The rst group of subsamples was stored at 4 ◦C to measure water
content (WC) and organic matter content (OM). The second group of
subsamples was dried at 65 ◦C for three days until a constant weight was
reached for measuring pH, electrical conductivity (EC), water extract-
able nutrients, total nutrients, and dissolved organic carbon (DOC). The
third group of subsamples was stored at −20 ◦C for DNA extraction.
2.2. Physicochemical analysis of soil samples
WC was measured after drying soil samples in a forced-air oven at
105 ◦C for 8 h. Afterward, samples were burned at 550 ◦C for 2 h to
quantify OM. Soil pH and EC were measured using a Mettler Toledo
SevenExcellence™ in a 1:10 soil/MilliQ water suspension (w/v) after 2 h
shaking at 25 ◦C. Water extractable nutrients, including NO
3
−
, NO
2
−
,
PO
4
3−
, NH
4
+
, and K
+
, were measured in a 1:10 soil/MilliQ water sus-
pension (w/v). The suspension was centrifuged at 3750 g after 2 h
shaking at 25 ◦C. The supernatant was ltered through a 0.45
μ
m
membrane lter and then analyzed by ion chromatography (Metrohm
Compact IC 761). Before analyzing total nutrients, soil samples were
ground into ne particles and mixed homogeneously. Total carbon (TC)
and total nitrogen (TN) were determined with an elemental analyzer
(Interscience FlashSmart CHNSO). Total phosphorus (TP) and total po-
tassium (TK) were analyzed by inductive coupled plasma optical emis-
sion spectrometry (Perkin Elmer Optima 5300 DV) after microwave
digestion with acids (Milestone Ethos Easy SK-15). DOC was extracted
and prepared the same way as water extractable nutrients, then
analyzed by a TOC analyzer (Shimadzu TOC-L). The effects of the OAs
and chemical fertilization on soil physicochemical properties were
tested by ANOVA (alpha =0.05), followed by pair-wise comparison
(TukeyHSD, family-wise error rate 5%). The assumptions of normality
and homogeneity of variances were assessed for each ANOVA model.
2.3. Microbial community analysis of soil
DNA was extracted from 0.25 g of soil using the DNeasy Power Soil
Kit (Qiagen) following the manufacturer’s protocol. DNA concentration
was quantied by Quantus (Promega), and purity was checked by a
Nanodrop spectrophotometer (Thermo Scientic NanoDrop 1000 spec-
trophotometer) with OD
260
and OD
280
. Extracted DNA samples were
stored at −20 ◦C. DNA samples were normalized to 20 ng/
μ
L for library
preparation and were sequenced (MrDNA, TX, USA) on a MiSeq (Illu-
mina). Libraries for bacteria were constructed using primers 338F
(ACTCCTACGGGAGGCAGCAG) (Fierer et al., 2005) and 806R (GGAC-
TACHVGGGTWTCTAAT) (Caporaso et al., 2011). Libraries for fungi
were constructed using primers ITS1F (CTTGGTCATTTA-
GAGGAAGTAA) and ITS2R (GCTGCGTTCTTCATCGATGC) (Adams
Table 1
Description of the eld experiment. Ten sub-treatments were clustered into three
treatments: unamended soil (control soil), compost amended soil, and bokashi
amended soil. Compost and bokashi had two application doses, and two levels of
chemical fertilization were applied as well. Each sub-treatment had three rep-
licates, resulting in 30 plots in total. The plots were randomly distributed. NA:
not applicable. F- and F+represent a low and a high application dose of N
fertilizer, respectively.
Treatment
Organic
amendment
type
Organic
amendment
dose (ton/ha)
N Fertilizer
dose (kg/
ha)
1 Control.
F- Unamended
soil (control)
NA NA 50
2 Control.
F+
NA NA 100
3 CL.F-
Compost
amended soil
Compost 3.8 50
4 CL.F+Compost 3.8 100
5 CH.F- Compost 9.5 50
6 CH.F+Compost 9.5 100
7 BL.F-
Bokashi
amended soil
Bokashi 9.7 50
8 BL.F+Bokashi 9.7 100
9 BH.F- Bokashi 24.25 50
10 BH.F+Bokashi 24.25 100
Y. Luo et al.
Soil Biology and Biochemistry 174 (2022) 108831
3
et al., 2013). The raw sequence data can be accessed at the European
Nucleotide Archive (ENA) at EMBL-EBI under accession number
PRJEB46153 (https://www.ebi.ac.uk/ena/browser/view/PRJEB46
153). Raw sequence data were analyzed by QIIME2 (version 2019.10)
following the methodology in our previous publication (Luo et al.,
2022). The downstream analyses of bacterial and fungal communities
were performed in RStudio (R version 4.2.1) using the phyloseq package
(McMurdie and Holmes, 2013) and the vegan package (Oksanen et al.,
2020).
2.4. Co-occurrence network analysis
Bacterial sequence data were rst quality trimmed and clustered (de
novo clustering) into OTUs with a 90% identity threshold using
VSEARCH (Rognes et al., 2016) in QIIME2. The 90% identity threshold
reduces the size of the OTU table signicantly compared with the more
standard “species-level” cutoff level of 97% identity, reducing the
network complexity and greatly facilitating visualization and analysis of
the data (Konstantinidis and Tiedje, 2007; Barber´
an et al., 2012). Fungal
sequence data were quality trimmed and clustered (closed-reference
clustering) into OTUs using a 97% identify threshold. After ltering (>
0.01% abundance, >4 samples prevalence) and clustering, we detected
275 bacterial OTUs (1,636,362 reads) and 270 fungal OTUs (1,591,059
reads) across all soil samples. Bacterial and fungal sequence data were
rst rareed at the same depth (29,073) and then concatenated into a
single OTU table. Core OTUs for network analysis were dened by
retaining OTUs detected in at least ve out of six replicate samples of a
treatment. Core OTUs comprised >88.2% of the total relative abun-
dance (Table 2). A total of 181 OTUs were shared across all treatments,
and the following OTU counts were retained per treatment: control (n =
244 OTUs), CL (n =255), CH (n =260), BL (n =260), BH (n =267).
Proportions of OTUs shared by and unique to the ve treatments are
presented as supplementary information (Fig. S1).
To determine if OA treatments differentially affected bacterial and
fungal co-occurrences, we used robust correlation analysis that accounts
for compositionality and sparsity using SparCC (Sparse Correlations for
Compositional data) (Friedman and Alm, 2012), implemented with the
SpiecEasi package (Kurtz et al., 2015). The correlation between two
OTUs was considered robust if the correlation coefcient >|0.6| and P
<0.01 (Barber´
an et al., 2012). Non-random co-occurrence patterns
were evaluated based on checkerboard scores (C-score) using a null
model approach that assumes no interactions between microbial taxa
(Stone and Roberts, 1990). C-scores were calculated using the vegan
package (Oksanen et al., 2020). A higher C-score denotes a higher level
of non-randomness or network structure. To assess network topologies,
we calculated the average degree, average clustering coefcient,
average path distance, and modularity. All calculations were performed
using the igraph packages (Csardi and Nepusz, 2006). Networks were
explored and visualized using Gephi (version 0.9.2) (Bastian et al.,
2009). Mantel tests were performed to estimate correlations between the
major network modules (considering OTUs in each module and their
relative abundance) and the physicochemical properties of the soil using
the ecodist package (Goslee and Urban, 2007).
3. Results
3.1. Soil physicochemical properties seven years after amendment
application
In general, the control soil and the CL soil had more similar physi-
cochemical properties than the other treatments. Specically, the con-
trol soil and the CL soil had lower concentrations of DOC and TC and
higher concentrations of TP and TK than the CH, BL, and BH soils
(Fig. 1). TN and other water extractable nutrients (i.e., AN, phosphate,
and potassium) were generally higher in the amended soils than the
control soils. OAs had no signicant effects on soil pH, EC, WC, and OM
(supplementary information (Table S1; Table S2)). The fertilizer dose
had no signicant impact on the physicochemical properties, except
phosphate (Table S1; Table S2). Therefore, we assessed the impact of
OAs and doses on the soil physicochemical variables without dis-
tinguishing between the two chemical fertilizer doses (Fig. 1).
3.2. Effects of OAs on co-occurrence networks of soil microorganisms
Chemical fertilization did not inuence the richness and composition
of the soil microbial community (Fig. S2; Table S3; Table S4). Therefore,
we assessed the effects of OAs and doses on soil microbial networks and
co-occurrence patterns without distinguishing between the two chemi-
cal fertilizer doses.
Robust correlation analysis revealed that OA amended soils retained
more nodes (OTUs) and edges than control soil (Fig. 2 and Table 2).
Higher C-scores were observed in the amended soils, indicating that the
addition of either compost or bokashi increased the network structure of
the soil microbial community. The control soil community network had
a C-score of 13.8 (p <0.001), which also indicated a non-random
structure. The addition of a low dose of bokashi (BL) resulted in the
highest C-score when all treatments were compared. The microbial taxa
(nodes) in the amended soils were densely connected into seemingly
more tightly knit groups than the nodes in the control soil, as indicated
by higher average degrees and higher average clustering coefcients.
In all treatments, soil microbial networks were structured by mod-
ules (modularity index >0.5; Table 2; Fig. 3). The modularity was
highest in control soil, which also had more modules than any of the
treatments (Table 2). The number of major modules (degree >5) across
all treatments was similar; however, the size (number of nodes) of major
modules in the amended soils was generally larger than in the control.
This was also reected by the average degree and average clustering
coefcient.
In addition, the formation of major modules in treatment-specic
networks was correlated differentially with the soil physicochemical
parameters (Fig. 4A). EC, DOC, TC, TN, AN, and phosphate were the
main soil physicochemical measures that signicantly correlated to
treatment-specic modules, but few signicant correlations (P <0.05)
were observed. The addition of compost and bokashi increased the
number of bacterial-fungal correlations compared with the unamended
soil, particularly in the BL treatment (Fig. 4B). This suggests that more
bacterial and fungal taxa in the amended soils co-occurred and poten-
tially interacted more frequently than in control soil. The addition of
compost and bokashi increased the proportion of positive correlations
while decreased the proportion of negative correlations among taxa
(Table 2 and Fig. 4B). Notably, the addition of compost and bokashi
decreased the proportions of negative correlations within kingdoms (i.
Table 2
Key properties of networks. Control: unamended soil; CL: soil amended with a
low dose of compost (3.8 ton/ha); CH: soil amended with a high dose of compost
(9.5 ton/ha); BL: soil amended with a low dose of bokashi (9.7 ton/ha); BH: soil
amended with a high dose of bokashi (24.25 ton/ha).
Control CL CH BL BH
No. of original nodes 244 255 260 260 267
total relative abundance 93.5% 96.8% 95.4% 95.9% 88.2%
After selection of robust correlation
No. of nodes 216 242 237 251 249
No. of edges 413 691 693 1104 741
positive edge/negative edge 0.39 0.63 1.24 0.54 0.67
edge density 0.02 0.02 0.02 0.04 0.02
C-score 13.8 29.9 30.0 66.4 31.7
average degree 3.8 5.7 5.8 8.8 6.0
average clustering coefcient 0.3 0.4 0.5 0.4 0.4
average path distance 5.5 5.0 5.9 3.8 4.9
modularity 0.77 0.70 0.72 0.53 0.65
No. of modules 16 11 11 7 11
major modules (degree >5) 9 9 7 7 9
Y. Luo et al.
Soil Biology and Biochemistry 174 (2022) 108831
4
e., bacteria-bacteria, fungi-fungi) compared to the unamended soil
(Fig. 4B). However, increased negative correlations across kingdoms (i.
e., bacteria-fungi) were observed in the amended soils, especially in the
BL treatment. Common taxa of bacteria and fungi were observed in
major modules (Fig. 4C). Taxa within Ascomycota fungi were highly
prevalent in the networks across all treatments. Other fungal taxa
(including Basidiomycota, Mortierellomycota, Glomeromycota, and
Chytridiomycota) were also observed in major modules. Diverse bacte-
rial phyla were also observed in major modules: Chloroexi, Proteo-
bacteria, Actinobacteria, Firmicutes, Bacteroidetes, Verrucomicrobia,
Acidobacteria, and Planctomycetes were dominant across all treatments.
4. Discussion
4.1. Non-random co-occurrence patterns induced by OAs
Non-random community assembly may be a general characteristic
across all life domains (Horner-Devine et al., 2007; Barber´
an et al.,
2012). Our study, which shows signicant non-random co-occurrence
patterns in the soil with and without the application of OAs, supports
this general notion. Our observation that the OA treatment increases the
complexity of networks in native soil is in line with previous studies
(Ling et al., 2016; Yang et al., 2019; Zhou et al., 2019). Our data support
that the microbial community in the amended soils was more organized
and potentially had a larger operational community that was function-
ally interrelated than that in control because of a higher C-score, more
edges, and more densely correlated nodes in the modules.
Interestingly, in our study, a decrease in the proportion of negative
correlations was found in the amended soils when compared to the
control soil, suggesting that amelioration of soil with compost and
bokashi may decrease competition among microorganisms for resources
(Blagodatskaya and Kuzyakov, 2008; Mau et al., 2015). Other studies
also reported that adding OAs or nutrients reduced the number and
strength of negative correlations between microorganisms (Banerjee
et al., 2016; Yang et al., 2019, 2020). The addition of OAs introduced
extra C and nutrients to the soil. This may reduce competition and in-
crease the number of multiple trophic levels or resource cascades in the
food web, which could improve energy use efciency (Coyte et al., 2015;
Banerjee et al., 2016). However, theoretical studies predict that
compartmentalization and the presence of negative interactions will
increase the stability of networks under disturbances, as negative links
stabilize co-oscillation in communities (Coyte et al., 2015; Kuiper et al.,
2015). In addition, a recent experimental study demonstrated that
negative interactions are more common than reciprocal positive in-
teractions in soil bacterial communities. Negative interactions can be
desirable by aiding in resistance against invasive species and main-
taining the stability of the community (Palmer and Foster, 2022). Even
though the proportion of negative correlations was reduced in the OA
amended soils, the absolute number of negative correlations was higher
than in the control soils. Therefore, the addition of compost or bokashi
may contribute to the stability of the microbial community by enhancing
resistance or resilience to disturbance.
4.2. Distinct modules observed in the amended soils
In all treatments, we observed high modularity indices, indicating
that the microbial community was structured into modules consisting of
groups of bacteria and fungi that formed strong associations. A modu-
larity index >0.4 suggests that the network can be well-divided into
modules or clusters with a strong modular structure (Newman, 2006).
These ndings support our hypothesis that the addition of different
doses of compost and bokashi will result in distinctive microbial mod-
ules in the co-occurrence networks. In addition, the size (number of
nodes) of major modules was larger in the amended soils than in the
control, particularly in the BL soil. This also suggests that the addition of
OAs may have changed the function of the soil microbial community
with larger operational communities that could be functionally
correlated.
Interestingly, only a few major modules signicantly correlated with
soil variables (i.e., EC, DOC, TC, TN, AN, and phosphate). Our results
Fig. 1. Effects of compost and bokashi on the physicochemical properties of soil seven years after application. Box plots are shown representing the median and
quartiles of the data. Experimental replication n =6 (without distinguishing the two levels of chemical fertilization). DOC: dissolved organic carbon; TC: total carbon;
TN: total nitrogen; TP: total phosphorus; TK: total potassium; AN: water extractable nitrogen, the sum of NO
3
−
, NO
2
−
, and NH
4
+
. Control: unamended soil; CL: soil
amended with a low dose of compost (3.8 ton/ha); CH: soil amended with a high dose of compost (9.5 ton/ha); BL: soil amended with a low dose of bokashi (9.7 ton/
ha); BH: soil amended with a high dose of bokashi (24.25 ton/ha).
Y. Luo et al.
Soil Biology and Biochemistry 174 (2022) 108831
5
suggest that the physicochemical properties of the soil might not be the
primary factors driving the formation of the network modules. We
speculate that the heavy clay soil in this study eld, characterized by
strong compaction (Roelfsema, 2010), may weaken the inuence of the
physicochemical properties of the soil on the microbial network. Alter-
natively, there may be unknown factors (e.g., specic and multiple in-
teractions with protists, Archaea, and other soil organisms) that are
more important than the physicochemical properties of the soil in
determining the soil microbial structure (Fuhrman, 2009; Deveau et al.,
2018). Further studies are suggested to examine the inuence of OAs on
microbial networks in different soil types (e.g., sand and clay) and how
the underground food web shapes microbial networks. Nevertheless, the
microbial network might be susceptible to changes in EC, DOC, TC, TN,
AN, and phosphate, which provide necessary nutrients for the growth,
activity, and metabolism of soil microorganisms.
4.3. Organic amendments changed intra- and inter-kingdom correlations
Bacteria and fungi operate in the same environment, where complex
interactions occur between these kingdoms, including predation, para-
sitism, competition, commensalism, and mutualism. Bacteria associated
with soil fungi, including saprotrophic and mycorrhizal ones, can benet
from nutritional interactions (Haq et al., 2014). Nutritional interactions
imply that bacteria feed on fungi or on their released products (De Boer
et al., 2005; Leveau and Preston, 2008) or vice versa (Hildebrandt et al.,
2002, 2006). Such nutritional interactions often imply commensalism or
true mutualism (Nazir et al., 2012). We expected that adding OAs would
weaken nutritional interactions between bacteria and fungi due to the
provisioning of additional nutrients; thus, fewer bacterial-fungal corre-
lations would be detected in the amended soils than in the unamended
soil. However, this was not in line with our observation. We observed
that the addition of compost and bokashi increased bacterial-fungal
Fig. 2. Network visualization of co-occurring bacteria and fungi per treatment. Control: unamended soil; CL: soil amended with a low dose of compost (3.8 ton/ha);
CH: soil amended with a high dose of compost (9.5 ton/ha); BL: soil amended with a low dose of bokashi (9.7 ton/ha); BH: soil amended with a high dose of bokashi
(24.25 ton/ha).
Y. Luo et al.
Soil Biology and Biochemistry 174 (2022) 108831
6
correlations, mainly increasing their negative correlations compared to
control, while positive correlations of bacteria-bacteria and fungi-fungi
were increased after applying OAs. We hypothesize that the provision-
ing of nutrients (TC and TN) and the availability of nutrients (DOC, AN,
and phosphate) by OAs may alleviate resource competition within
kingdoms but stimulate competition across kingdoms. It is unclear what
drives this. Even though efforts have been taken to uncover mechanisms
driving bacterial-fungal interactions; challenges remain since the
outcome of these interactions are the combined results of the physical
associations (biolm, free cell), molecular communication between
(micro)organisms, and the local soil environment (Deveau et al., 2018).
Further research should identify essential parameters in the soil system
that can drive bacterial-fungal interactions.
We emphasize that the intra and inter-kingdom co-occurrence pat-
terns are correlations between bacterial and fungal taxa, which indicates
potential positive, negative, or neutral interactions but cannot uncover
causal mechanisms structuring the communities (Barber´
an et al., 2012;
Banerjee et al., 2016). Moreover, correlation analyses inherently imply
technical challenges when analyzing microbial associations in soil, such
as how to pre-process compositional abundance data (Faust, 2021). In
addition, the performance of the statistical methods (including simple,
direct correlation measures like Pearson and Spearman correlations) for
inferring microbial association does not account for data properties that
may yield spurious associations (Weiss et al., 2016; Hirano and Take-
moto, 2019; Faust, 2021). Spearman correlation is still commonly
applied to infer microbial networks from non-independent relative
abundance data, but negative correlations are often overrepresented
regardless of the true correlations underlying absolute abundances
(Friedman and Alm, 2012). Methodological choices for network ana-
lyses need careful consideration of the data characteristics. Particularly,
correlation techniques should be considered that account for sparsity
and compositionality when necessary, such as SparCC (Friedman and
Fig. 3. Major modules (degree >5) highlighted in networks per treatment. Control: unamended soil; CL: soil amended with a low dose of compost (3.8 ton/ha); CH:
soil amended with a high dose of compost (9.5 ton/ha); BL: soil amended with a low dose of bokashi (9.7 ton/ha); BH: soil amended with a high dose of bokashi
(24.25 ton/ha).
Y. Luo et al.
Soil Biology and Biochemistry 174 (2022) 108831
7
Alm, 2012). Nonetheless, our results highlight the importance of
considering prokaryotic and eukaryotic components of the soil micro-
bial community, which is often neglected when microbial communities
are investigated as compartmentalized groups (De Menezes et al., 2015).
5. Conclusion
Compost and bokashi changed the intra- and inter-kingdom co-
occurrence patterns of microorganisms in soils. Soil microbial commu-
nities appeared to have a higher level of non-randomness in the net-
works after applying compost and bokashi. This was particularly so
when bokashi was applied at a low dose. OAs drove distinct network
modules in terms of module compositions and the size of major modules.
Our results suggest that network modules correlated with soil recourses
(i.e., EC, DOC, TC, TN, AN, and phosphate). We also observed that OAs
decreased the proportion of negative correlations, reducing potential
competition among microorganisms for resources and increasing energy
use efciency. The addition of OAs increased bacterial-fungal correla-
tions, particularly their negative correlations. Further studies should be
conducted to identify key factors driving the changes in bacterial-fungal
interactions after OAs application. In addition, it is essential to have
experimental evidence validating the hypotheses generated by micro-
bial networks (e.g., how abiotic properties of soil shape biotic correla-
tions) since network analyses do not uncover causal relationships. The
links between networks and soil ecological functions (e.g., plant growth,
C sequestration, and soil aggregation) also merit further research.
Declaration of competing interest
The authors declare that they have no known competing nancial
interests or personal relationships that could have appeared to inuence
the work reported in this paper.
Data availability
Data will be made available on request.
Acknowledgments
This work was performed in the cooperation framework of Wetsus,
European Centre of Excellence for Sustainable Water Technology (www.
wetsus.eu). Wetsus is co-funded by the Dutch Ministry of Economic
Affairs and Climate Policy, the European Union Regional Development
Fund, the City of Leeuwarden, the Province of Fryslˆ
an, the Northern
Netherlands Provinces, and the Netherlands Organisation for Scientic
Research. The authors would also like to thank the members of the
research soil theme (Agriton, Mulder Agro, Netherlands Institute of
Ecology (NIOO-KNAW), Koninklijke Oosterhof Holman, Waterketen
Onderzoek Noord (WON), and Waterschap Zuiderzeeland) for the
fruitful discussions and nancial support. The authors are grateful to the
Fig. 4. (A) Correlation between major modules and the physicochemical properties of the soil based on the Mantel test. Signicant correlations (P <0.05) were
marked with “X”; (B) Frequency of intra- and inter-kingdom correlations in the networks. The proposition of negative and positive correlation within each type of
correlation is indicated by different color intensities: the dark shade of the color represents negative correlations, and the light shade represents positive correlations;
(C) Frequency of nodes (bacteria and fungi) at phylum level in major modules detected in the networks. Control: unamended soil; CL: soil amended with a low dose of
compost (3.8 ton/ha); CH: soil amended with a high dose of compost (9.5 ton/ha); BL: soil amended with a low dose of bokashi (9.7 ton/ha); BH: soil amended with a
high dose of bokashi (24.25 ton/ha). (For interpretation of the references to color in this gure legend, the reader is referred to the Web version of this article.)
Y. Luo et al.
Soil Biology and Biochemistry 174 (2022) 108831
8
editor and reviewers for their time and valuable comments.
Appendix A. Supplementary data
Supplementary data to this article can be found online at https://doi.
org/10.1016/j.soilbio.2022.108831.
References
Adams, R.I., Miletto, M., Taylor, J.W., Bruns, T.D., 2013. Dispersal in microbes: fungi in
indoor air are dominated by outdoor air and show dispersal limitation at short
distances. The ISME Journal 7, 1262–1273.
Banerjee, S., Kirkby, C.A., Schmutter, D., Bissett, A., Kirkegaard, J.A., Richardson, A.E.,
2016. Network analysis reveals functional redundancy and keystone taxa amongst
bacterial and fungal communities during organic matter decomposition in an arable
soil. Soil Biology and Biochemistry 97, 188–198.
Barber´
an, A., Bates, S.T., Casamayor, E.O., Fierer, N., 2012. Using network analysis to
explore co-occurrence patterns in soil microbial communities. The ISME Journal 6,
343–351.
Bastian, M., Heymann, S., Jacomy, M., 2009. Gephi: an open source software for
exploring and manipulating networks. International AAAI Conference on Weblogs
and Social Media 8, 361–362.
Bezemer, T.M., Fountain, M.T., Barea, J.M., Christensen, S., Dekker, S.C., Duyts, H., Van
Hal, R., Harvey, J.A., Hedlund, K., Maraun, M., Mikola, J., Mladenov, A.G.,
Robin, C., De Ruiter, P.C., Scheu, S., Set¨
al¨
a, H., ˇ
Smilauer, P., Van Der Putten, W.H.,
2010. Divergent composition but similar function of soil food webs of individual
plants: plant species and community effects. Ecology 91, 3027–3036.
Blagodatskaya, E., Kuzyakov, Y., 2008. Mechanisms of real and apparent priming effects
and their dependence on soil microbial biomass and community structure: critical
review. Biology and Fertility of Soils 45, 115–131.
Bronswijk, J.J.B., Evers-Vermeer, J.J., 1990. Shrinkage of Dutch clay soil aggregates.
Netherlands Journal of Agricultural Science 38, 175–194.
Caporaso, J.G., Lauber, C.L., Walters, W.A., Berg-Lyons, D., Lozupone, C.A.,
Turnbaugh, P.J., Fierer, N., Knight, R., 2011. Global patterns of 16S rRNA diversity
at a depth of millions of sequences per sample. Proceedings of the National Academy
of Sciences of the United States of America 108, 4516–4522.
Coyte, K.Z., Schluter, J., Foster, K.R., 2015. The ecology of the microbiome: networks,
competition, and stability. Science 350, 663–666.
Csardi, G., Nepusz, T., 2006. The igraph software package for complex network research.
InterJournal, Complex Systems 1695, 1–9.
De Boer, W., Folman, L.B., Summerbell, R.C., Boddy, L., 2005. Living in a fungal world:
impact of fungi on soil bacterial niche development. FEMS Microbiology Reviews 29,
795–811.
De Menezes, A.B., Prendergast-Miller, M.T., Richardson, A.E., Toscas, P., Farrell, M.,
Macdonald, L.M., Baker, G., Wark, T., Thrall, P.H., 2015. Network analysis reveals
that bacteria and fungi form modules that correlate independently with soil
parameters. Environmental Microbiology 17, 2677–2689.
Deveau, A., Bonito, G., Uehling, J., Paoletti, M., Becker, M., Bindschedler, S.,
Hacquard, S., Herv´
e, V., Labb´
e, J., Lastovetsky, O.A., Mieszkin, S., Millet, L.J.,
Vajna, B., Junier, P., Bonfante, P., Krom, B.P., Olsson, S., van Elsas, J.D., Wick, L.Y.,
2018. Bacterial-fungal interactions: ecology, mechanisms and challenges. FEMS
Microbiology Reviews 42, 335–352.
Dumontet, S., Cavoski, I., Ricciuti, P., Mondelli, D., Jarrar, M., Pasquale, V., Crecchio, C.,
2017. Metabolic and genetic patterns of soil microbial communities in response to
different amendments under organic farming system. Geoderma 296, 79–85.
Faust, K., 2021. Open challenges for microbial network construction and analysis. The
ISME Journal 15, 3111–3118.
Fierer, N., Jackson, J.A., Vilgalys, R., Jackson, R.B., 2005. Assessment of soil microbial
community structure by use of taxon-specic quantitative PCR assays. Applied and
Environmental Microbiology 71, 4117–4120.
Friedman, J., Alm, E.J., 2012. Inferring correlation networks from genomic survey data.
PLoS Computational Biology 8, 1–11.
Fritsche, K., Van den Berg, M., De Boer, W., Van Beek, T.A., Raaijmakers, J.M., Van
Veen, J.A., Leveau, J.H.J., 2014. Biosynthetic genes and activity spectrum of
antifungal polyynes from Collimonas fungivorans Ter331. Environmental
Microbiology 16, 1334–1345.
Fuhrman, J.A., 2009. Microbial community structure and its functional implications.
Nature 459, 193–199.
Goslee, S.C., Urban, D.L., 2007. The ecodist package for dissimilarity-based analysis of
ecological data. Journal of Statistical Software 22, 1–19.
Haq, I.U., Zhang, M., Yang, P., Van Elsas, J.D., 2014. The interactions of bacteria with
fungi in soil: emerging concepts. Advances in Applied Microbiology 89, 185–215.
Hildebrandt, U., Janetta, K., Bothe, H., 2002. Towards growth of arbuscular mycorrhizal
fungi independent of a plant host. Applied and Environmental Microbiology 68,
1919–1924.
Hildebrandt, U., Ouziad, F., Marner, F.J., Bothe, H., 2006. The bacterium Paenibacillus
validus stimulates growth of the arbuscular mycorrhizal fungus Glomus intraradices
up to the formation of fertile spores. FEMS Microbiology Letters 254, 258–267.
Hirano, H., Takemoto, K., 2019. Difculty in inferring microbial community structure
based on co-occurrence network approaches. BMC Bioinformatics 20, 1–14.
Horner-Devine, M.C., Silver, J.M., Leibold, M.A., Bohannan, B.J.M., Colwell, R.K.,
Fuhrman, J.A., Green, J.L., Kuske, C.R., Martiny, J.B.H., Muyzer, G., Øvreås, L.,
Reysenbach, A.L., Smith, V.H., 2007. A comparison of taxon co-occurrence patterns
for macro- and microorganisms. Ecology 88, 1345–1353.
Jenkinson, D.S., 1977. The soil biomass. New Zealand Soil News 25, 213–218.
Konstantinidis, K.T., Tiedje, J.M., 2007. Prokaryotic taxonomy and phylogeny in the
genomic era: advancements and challenges ahead. Current Opinion in Microbiology
10, 504–509.
Kuiper, J.J., Van Altena, C., De Ruiter, P.C., Van Gerven, L.P.A., Janse, J.H., Mooij, W.M.,
2015. Food-web stability signals critical transitions in temperate shallow lakes.
Nature Communications 6, 1–7.
Kurtz, Z.D., Müller, C.L., Miraldi, E.R., Littman, D.R., Blaser, M.J., Bonneau, R.A., 2015.
Sparse and compositionally robust inference of microbial ecological networks. PLoS
Computational Biology 11, 1–25.
Leveau, J.H.J., Preston, G.M., 2008. Bacterial mycophagy: denition and diagnosis of a
unique bacterial-fungal interaction. New Phytologist 177, 859–876.
Ling, N., Zhu, C., Xue, C., Chen, H., Duan, Y., Peng, C., Guo, S., Shen, Q., 2016. Insight
into how organic amendments can shape the soil microbiome in long-term eld
experiments as revealed by network analysis. Soil Biology and Biochemistry 99,
137–149.
Lu, H., Lashari, M.S., Liu, X., Ji, H., Li, L., Zheng, J., Kibue, G.W., Joseph, S., Pan, G.,
2015. Changes in soil microbial community structure and enzyme activity with
amendment of biochar-manure compost and pyroligneous solution in a saline soil
from Central China. European Journal of Soil Biology 70, 67–76.
Lu, L., Yin, S., Liu, X., Zhang, W., Gu, T., Shen, Q., Qiu, H., 2013. Fungal networks in
yield-invigorating and -debilitating soils induced by prolonged potato monoculture.
Soil Biology and Biochemistry 65, 186–194.
Luo, Y., van Veelen, H.P.J., Chen, S., Sechi, V., ter Heijne, A., Veeken, A., Buisman, C.J.
N., Bezemer, T.M., 2022. Effects of sterilization and maturity of compost on soil
bacterial and fungal communities and wheat growth. Geoderma 409, 115598.
Ma, B., Wang, H., Dsouza, M., Lou, J., He, Y., Dai, Z., Brookes, P.C., Xu, J., Gilbert, J.A.,
2016. Geographic patterns of co-occurrence network topological features for soil
microbiota at continental scale in eastern China. The ISME Journal 10, 1891–1901.
Mau, R.L., Liu, C.M., Aziz, M., Schwartz, E., Dijkstra, P., Marks, J.C., Price, L.B., Keim, P.,
Hungate, B.A., 2015. Linking soil bacterial biodiversity and soil carbon stability. The
ISME Journal 9, 1477–1480.
McMurdie, P.J., Holmes, S., 2013. Phyloseq: an R package for reproducible interactive
analysis and graphics of microbiome census data. PLoS One 8, e61217.
Nazir, R., Zhang, M., de Boer, W., van Elsas, J.D., 2012. The capacity to comigrate with
Lyophyllum sp. strain Karsten through different soils is spread among several
phylogenetic groups within the genus Burkholderia. Soil Biology and Biochemistry
50, 221–233.
Neher, D.A., Weicht, T.R., Bates, S.T., Leff, J.W., Fierer, N., 2015. Changes in bacterial
and fungal communities across compost recipes, preparation methods, and
composting times. Biological Treatment of Solid Waste: Enhancing Sustainability 8,
119–144.
Newman, M.E.J., 2006. Modularity and community structure in networks. In:
Proceedings of the National Academy of Sciences of the United States of America,
vol. 103, pp. 8577–8582.
Oksanen, J., Blanchet, F., Friendly, M., Kindt, R., Legendre, P., McGlinn, D., Minchin, P.,
O’Hara, R., Simpson, G., Solymos, P., Stevens, M., Szoecs, E., Wagner, H., 2020.
Vegan: community ecology package. R Package Version 2, 5–7.
Palmer, J.D., Foster, K.R., 2022. Bacterial species rarely work together. Science 376,
581–582.
P´
erez-Piqueres, A., Edel-Hermann, V., Alabouvette, C., Steinberg, C., 2006. Response of
soil microbial communities to compost amendments. Soil Biology and Biochemistry
38, 460–470.
Prosser, J.I., Bohannan, B.J.M., Curtis, T.P., Ellis, R.J., Firestone, M.K., Freckleton, R.P.,
Green, J.L., Green, L.E., Killham, K., Lennon, J.J., Osborn, A.M., Solan, M., Gast, C.J.
van der, Young, J.P.W., 2007. The role of ecological theory. Nature Reviews
Microbiology 5, 384–392.
Roelfsema, E., 2010. Inuence of the Shrink and Swell Cycle on Soil Physical Properties
of Loam and Clay Soils in the Netherlands. Master Thesis. Wageningen University &
Research. https://edepot.wur.nl/160236.
Rognes, T., Flouri, T., Nichols, B., Quince, C., Mah´
e, F., 2016. VSEARCH: a versatile open
source tool for metagenomics. PeerJ 4, e2584.
Schmid, C.A.O., Schr¨
oder, P., Armbruster, M., Schloter, M., 2018. Organic amendments
in a long-term eld trial—consequences for the bulk soil bacterial community as
revealed by network analysis. Microbial Ecology 76, 226–239.
Shin, K., van Diepen, G., Blok, W., van Bruggen, A.H.C., 2017. Variability of effective
micro-organisms (EM) in bokashi and soil and effects on soil-borne plant pathogens.
Crop Protection 99, 168–176.
Sparling, G.P., 1985. The soil bomass. Soil Organic Matter and Biological Activity 25,
223–262.
Stone, L., Roberts, A., 1990. The checkerboard score and species distributions. Oecologia
85, 74–79.
Wagg, C., Schlaeppi, K., Banerjee, S., Kuramae, E.E., van der Heijden, M.G.A., 2019.
Fungal-bacterial diversity and microbiome complexity predict ecosystem
functioning. Nature Communications 10, 1–10.
Weiss, S., Van Treuren, W., Lozupone, C., Faust, K., Friedman, J., Deng, Y., Xia, L.C.,
Xu, Z.Z., Ursell, L., Alm, E.J., Birmingham, A., Cram, J.A., Fuhrman, J.A., Raes, J.,
Sun, F., Zhou, J., Knight, R., 2016. Correlation detection strategies in microbial data
sets vary widely in sensitivity and precision. The ISME Journal 10, 1669–1681.
Xue, C., Ryan Penton, C., Zhu, C., Chen, H., Duan, Y., Peng, C., Guo, S., Ling, N., Shen, Q.,
2018. Alterations in soil fungal community composition and network assemblage
structure by different long-term fertilization regimes are correlated to the soil
ionome. Biology and Fertility of Soils 54, 95–106.
Y. Luo et al.
Soil Biology and Biochemistry 174 (2022) 108831
9
Yang, W., Jing, X., Guan, Y., Zhai, C., Wang, T., Shi, D., Sun, W., Gu, S., 2019. Response
of fungal communities and co-occurrence network patterns to compost amendment
in black soil of northeast China. Frontiers in Microbiology 10, 1562.
Yang, W., Yang, Z., Guan, Y., Zhai, C., Shi, D., Chen, J., Wang, T., Gu, S., 2020. Dose-
dependent effect of compost amendment on soil bacterial community composition
and co-occurrence network patterns in soybean agroecosystem. Archives of
Agronomy and Soil Science 66, 1027–1041.
Zhou, Z., Gao, T., Zhu, Q., Yan, T., Li, D., Xue, J., Wu, Y., 2019. Increases in bacterial
community network complexity induced by biochar-based fertilizer amendments to
karst calcareous soil. Geoderma 337, 691–700.
Y. Luo et al.
