- Open Access
Protist communities are more sensitive to nitrogen fertilization than other microorganisms in diverse agricultural soils
© The Author(s). 2019
- Received: 8 November 2018
- Accepted: 13 February 2019
- Published: 27 February 2019
Agricultural food production is at the base of food and fodder, with fertilization having fundamentally and continuously increased crop yield over the last decades. The performance of crops is intimately tied to their microbiome as they together form holobionts. The importance of the microbiome for plant performance is, however, notoriously ignored in agricultural systems as fertilization disconnects the dependency of plants for often plant-beneficial microbial processes. Moreover, we lack a holistic understanding of how fertilization regimes affect the soil microbiome. Here, we examined the effect of a 2-year fertilization regime (no nitrogen fertilization control, nitrogen fertilization, and nitrogen fertilization plus straw amendment) on entire soil microbiomes (bacteria, fungi, and protist) in three common agricultural soil types cropped with maize in two seasons.
We found that the application of nitrogen fertilizers more strongly affected protist than bacterial and fungal communities. Nitrogen fertilization indirectly reduced protist diversity through changing abiotic properties and bacterial and fungal communities which differed between soil types and sampling seasons. Nitrogen fertilizer plus straw amendment had greater effects on soil physicochemical properties and microbiome diversity than nitrogen addition alone. Moreover, nitrogen fertilization, even more together with straw, increased soil microbiome network complexity, suggesting that the application of nitrogen fertilizers tightened soil microbiomes interactions.
Together, our results suggest that protists are the most susceptible microbiome component to the application of nitrogen fertilizers. As protist communities also exhibit the strongest seasonal dynamics, they serve as the most sensitive bioindicators of soil changes. Changes in protist communities might have long-term effects if some of the key protist hubs that govern microbiome complexities as top microbiome predators are altered. This study serves as the stepping stone to promote protists as promising agents in targeted microbiome engineering to help in reducing the dependency on exogenous unsustainably high fertilization and pesticide applications.
- Soil protists
- Soil type
- Nitrogen fertilizers
- High-throughput sequencing
Soils provide the basis of our life. We not only live on soils, but soils are also needed for crop production that forms the base of most of our food . While the importance of the abiotic part of soils including soil water and nutrient levels is widely acknowledged, the role of soil biota has received less attention. In fact, soils serve as the habitat of arguably most of the taxonomic biodiversity on planet Earth. We are slowly beginning to decipher this vast biodiversity and to understand the functional importance of the crop-associated microbiome especially in the plant-influenced rhizosphere in promoting plant growth, increasing plant tolerance, and protecting plants against pathogens [2–7]. Despite the acknowledged importance of the soil biodiversity in sustaining soil fertility and controlling plant performance, the positive functional role of soil biodiversity especially in agricultural systems is often ignored or treated as predominantly being plant deleterious that needs to be controlled. Artificial pesticides and fertilizers have disconnected our dependency on positive services soil biodiversity provides for plant performance, but long-term pesticide and fertilizer applications are increasingly shown to be unsustainable . To increase sustainability, we first need a better understanding of the processes induced by agricultural managements on soil biodiversity.
Nitrogen fertilization is a commonly used agricultural practice to increase crop production . However, the intensive application can lead to soil acidification, greenhouse gas N2O emission, nitrate leaching [10, 11], and biodiversity losses [12, 13]. Biodiversity decreases subsequently affect nitrogen cycling and soil multifunctionality [2, 14, 15]. As a traditional agronomic practice to improve soil fertility, straw retention showed significant effects on improving soil organic carbon contents and at reducing nitrogen loss [16, 17]. The application of combined organic and chemical nitrogen fertilizers is thus encouraged in modern sustainable and environmental-friendly agriculture. However, to date, the effects of nitrogen addition alone or in combination with straw incorporation on entire soil microbiomes are not understood in detail.
Most work examining the impact of fertilization on microbiome composition and function has focused on bacterial and fungal communities. Their community diversity was shown to be reduced by long-term nitrogen fertilizer application [9, 13, 18]. Protists as another microbial group have received little attention despite their key role in controlling bacterial and fungal populations . Previous morphogroup-based studies focusing on amoebae have shown that both nitrogen addition and plant residue amendment could affect protist density, while plant residue did not influence the community structure of protists [20, 21]. However, the taxonomic resolution of these morphogroup-approaches is low, with individual morphogroups consisting of phylogenetically and functionally diverse taxa  and might therefore undermine important ecological differences. High-throughput sequencing approaches now enable a deeper resolution of protist communities and have revealed that long-term mineral fertilization reduced protist diversity in soils compared with organic fertilization . In contrast, Lentendu et al.  reported that organic fertilizers more strongly altered protist communities than mineral fertilizers. The difference between these studies might have originated from a focus on single soil types and sampling points as protist communities are structured by abiotic conditions [24–28] that often differ between soil types and can vary over sampling times . Therefore, we miss a cumulative understanding of how fertilization, particularly widely applied nitrogen amendments, affects soil protists. Furthermore, to better understand the entire microbiome responses to fertilizer application and to disentangle the role of key microbial taxa in microbiome communities after fertilization, an integrative study of all microbial groups is needed.
We here performed a holistic analysis of soil microbiome members including bacteria, fungi, and protists using group-specific high-throughput sequencing approaches in three major agricultural crop production areas, corresponding to three major soil types, in China. These soils were planted with the same crop species (Zea mays) and treated with a consistent fertilization regime over 2 years. The microbiome structure was analyzed in summer and autumn. Our first hypothesis was that the application of nitrogen fertilizers reduces soil microbial diversity, with more pronounced effects on first trophic level bacteria and fungi than the predominantly predatory protists. Furthermore, we expected that straw amendments could mitigate the negative nitrogen-fertilizer-induced effects on biodiversity (hypothesis 2). We also tested whether these effects differ between soil types and sampling seasons. We analyzed diverse abiotic parameters to disentangle the importance of abiotic and biotic factors regulating microbiome interactions through structural equation models (SEM) and network analyses.
Soil physicochemical properties
The effects of soil type, season, and fertilization treatment on the differentiations of soil physicochemical properties and bacterial, fungal, and protist communities based on PERMANOVA
Fertilization × soil type
Fertilization × season
Soil type × season
Fertilization × soil type × season
The application of nitrogen fertilizers (N and NS treatments) significantly influenced the overall soil physicochemical properties (synthesized by pH, moisture, C/N ratio, organic matter, dissolved organic carbon, total content of carbon and nitrogen, ammonium nitrogen, and nitrate nitrogen; P = 0.001, R2 = 0.063; Table 1). The N and NS treatments increased soil NH4+-N and NO3−-N (Additional file 1: Table S2). However, the effects of fertilization on abiotics differed between soil types. In comparison to the control treatment, the NS treatment significantly decreased pH (P < 0.05) in the black and red soils (Additional file 1: Table S2). In addition, the NS treatment significantly increased NO3−-N (P < 0.05) in the red soil (Additional file 1: Table S2). The effects of fertilization also differed between seasons. In comparison to the control treatment, the N treatment significantly (P < 0.05) decreased pH in the red soil in summer and in the black soil in autumn (Additional file 1: Table S2). The N treatment significantly increased NO3−-N (P < 0.05) in the black and red soils in summer and increased NH4+-N (P < 0.05) in the black soil in autumn (Additional file 1: Table S2). The NS treatment significantly increased NH4+-N (P < 0.05) in the fluvo-aquic soil in summer and in the black and red soils in autumn (Additional file 1: Table S2). Generally, the NS treatment induced the strongest changes in soil physicochemical properties (Additional file 1: Table S2) such as leading to higher contents of NH4+-N and NO3−-N, while reducing pH in soil (Table 3).
The effects of soil type, season, and fertilization treatment on the changes of alpha diversity of bacterial, fungal, and protist communities based on linear mixed model (LMM)
Fertilization × soil type
Fertilization × season
Soil type × season
Fertilization × soil type × season
Overview of the significance difference based on one-way ANOVA for the effect of nitrogen fertilizers on soil physicochemical properties and alpha diversity of microbiomes in comparison to the control treatment
After quality filtering, the remaining 1,998,108 sequences were clustered into 519 fungal OTUs. The alpha diversity of fungal communities was not affected (P > 0.05) by the application of nitrogen fertilizers (N and NS treatments), except in the black and red soils in summer, in which the NS treatment (P < 0.05) significantly suppressed the alpha diversity of fungal communities (Fig. 1B, Table 3). Furthermore, there was no difference (P > 0.05) in the alpha diversity of fungal community between summer and autumn, except for the control treatment (P < 0.05) in the red soil (Fig. 1B). Similar to bacterial community, the application of nitrogen fertilizers (R2 = 0.017) had weaker effects than soil types (R2 = 0.369) on fungal community composition and diversity (Tables 1 and 2, Fig. 1E). LEfSe analysis revealed that 51 biomarkers affiliating with five phyla were sensitive to nitrogen fertilizer treatments in soils (P < 0.05, LDA > 2.0; Additional file 1: Figure S3, Table S4). These biomarkers were equal to 9.53% of all taxa retrieved. Unlike bacteria, the relative abundance of fungal taxa significantly changed (P < 0.05, LDA > 2.0) at a higher taxonomic level under nitrogen fertilization treatments. The phylum Ascomycota was significantly increased in the NS treatment while the phyla Glomeromycota and Chytridiomycota significantly decreased in autumn in the N and NS treatments in the black and fluvo-aquic soil, respectively (Additional file 1: Figure S3, Table S4).
Co-occurrence between soil microbiomes
Topological indices of each network in Fig. 3
All soil types
Number of nodes
Number of edges
Protist nodes (%)
Fungal nodes (%)
Bacterial nodes (%)
Edges linking protist to bacteria (%)
Edges linking protist to fungi (%)
Edges linking bacteria to fungi (%)
Protist nodes sensitively responding to nitrogen fertilization in each soil type
Further results for seasonal dynamics of soil microbiome networks are provided as Additional file 2.
Influential factors on soil microbiome communities
We here show that the application of nitrogen fertilizers affects the composition of the entire microbiome, particularly reducing soil protist diversity, with differences depending on plant growing season and soil type. We reveal that not only abiotic differences, but direct interactions within the microbiome drive the observed changes, with protists appearing as ecologically relevant microbiome components due to their predation manner on other microorganisms [30–32]. Protists were also most sensitive in the microbiome as their communities changed most strongly between fertilization treatments and seasons. Together with their small size, enormous diversity, high abundances, and broader presence in all environments , our results enforced the potential of protist communities as promising indicators that should be prioritized in future efforts to be implemented for, e.g., biomonitoring and pesticide indication. As protists were determined as key hubs in microbiome networks, changes in protist community might substantially influence soil microbiome associations through the top-control forces which need to be followed on.
Abiotic changes induced by fertilization were detected to structure microbiomes that were dependent on soil type and season. Numerous studies have demonstrated that long-term inorganic nitrogen fertilization changes soil bacterial and fungal communities through fertilization-induced changes in soil abiotics [9, 34, 35]. In this study, nitrogen fertilizer amendments significantly influenced the bacterial and fungal community composition but did not influence their diversity, while the effect was much weaker compared with that on the protist community. The minor or non-significant change in pH (less than 0.42) and other abiotics under a relatively short-term fertilizer application period may partially explain the weak effect of nitrogen fertilization on bacterial and fungal community, as fertilization-induced abiotic changes might only emerge after years [36–38]. On the other side, protists might have a lower habitat niche breadths than bacteria , suggesting that protists are less tolerant than bacteria to environmental changes. All these results were in contrast to our first hypothesis. Previous studies revealed inconsistent conclusions on how nitrogen fertilization affects soil protist communities [23, 40]. We here reveal that short-term application of nitrogen fertilizers reduces protist diversity. In contrast to our hypothesis 2, straw amendments further lowered protist diversity, which could be attributed to the increased nitrogen retention under the straw incorporation in soil as demonstrated by 15N-urea tracing in our trial field in the black soil  and other studies [41, 42]. Based on the data, we cannot decipher the specific underlying mechanisms of protist diversity decreases, such as through direct effects or via changes in abiotic or biotic components. However, some competitively strong opportunistic taxa could be promoted that outcompete and therefore simplify protist communities . As the loss of protist diversity induced by nitrogen fertilizers could subsequently trigger negative impacts on soil ecosystem stability and soil multifunctionality , it seems pivotal to better understand and manage soil microbiomes for a more sustainable agriculture. This integrative knowledge on microbiome functioning is, however, only achieved when the entire microbiome and not only individual parts are studied . Therefore, investigations on potential soil microbiome interactions such as in this study are urgently needed.
In general, abiotics profoundly contributed to reductions in the alpha diversity of the soil protist community. As bacteria and fungi are a major prey for phagotrophic protists in soil [44–46], biotic interactions within the soil microbiome could also lead to reductions in protist diversity following fertilizations through bottom-up driving. However, bottom-up driving forces of bacteria and fungi on protists differed between soil types, which are explained by differences in moisture and C/N. The red soil site located in south China has much more rainfall than the other two sites in north China; thus, the red soil often contains higher moisture which makes it easy for protists to feed on their bacterial prey in water-filled soil pores . Besides, fungi are more dominant in soil with higher C/N (like in the black and fluvo-aquic soils) than bacteria . Therefore, the bacterial and fungal community differs in their effects on the protist diversity depending on soil abiotic properties.
Soil microbiome networks visualized the scenarios of biotic interactions and indicated that protist taxa are tightly linked within microbiomes as potential key microbiome controllers. Specific and different links between protist taxa and microbial taxa indicate that protist in these families feed in a taxon-specific manner on their potentially microbial prey. This confirms previous studies showing that protist taxa selectively feed on bacterial  and fungal prey . Interestingly, however, we found variable links of some protist groups with different members of the microbiome in distinct soil types, seasons, and fertilization regimes. This suggests that these groups shift their predominant feeding and consequently the top-down impact on microbiome composition .
We found that the application of nitrogen fertilizers tightened soil microbiome association. This might be explained by changes of some taxa that were sensitive to nitrogen addition. The LEfSe analysis detected several nitrogen-susceptible taxa in the Rhizobiales, Xanthomonadales, and Burkholderiales which have previously been reported to be keystone taxa in agricultural ecosystems linked to rhizosphere function . The application of nitrogen fertilizers reduced protist diversity and altered the links of key protist taxa in the microbiome networks, which also possibly contributed to the tightening of soil microbiome association. Alternatively, negative impacts of high nitrogen availability on soil microbiomes  could promote few strong associations with tightened microbiome complexity and consequently increase system stability to resist to adverse environmental conditions [48–50]. Lastly, the exogenous nutrient addition increased resource and food availability and subsequently could strengthen microbial interactions in order to enhance the efficiency of resource turnover that benefits plant growth [48, 49, 51]. Generally, the application of nitrogen fertilizers has tightened soil microbiome associations. Mechanisms that explain the observed differences in the microbiomes of different soil types yet need to be determined. Further results and discussions such as networks specific for sampling season are provided as supplementary information. Taken together, we propose that protists may be keystone of soil microbiome as protists strongly correlated and therefore potentially drive microbiome structure.
The study provides a vital scenario on the dynamics of how the soil microbiome responds to fertilization under a relatively short-term scale of two years and is the first to demonstrate that the application of nitrogen fertilizers is reducing protist community diversity and is tightening soil microbiome network associations in diverse agricultural soils. However, these findings are based on a 2-year field trial and need to be tested in other systems, with other crops, and on different continents. Besides, our results reveal that changes in abiotic parameters differed between soil types and varied across sampling season. This indicates that more repetitive samplings across broader temporal and spatial scales are needed to accurately depict abiotic parameters that determine soil microbiomes, as abiotic parameters might be changed further after long-term fertilization regime. Therefore, controlled experimental studies are needed to disentangle the main abiotic and biotic drivers of protist communities as well as that of entire soil microbiomes.
Taken together, we show that the application of nitrogen fertilizers has a profound impact on soil microbiomes that differs between seasons and soil types, suggesting that studies focusing on a single soil type or sampling point cannot capture microbiome fluctuations that profoundly change microbiome composition throughout time. While fertilization also altered the bacterial and fungal community composition without affecting the diversity of these microbial groups, fertilization much more strongly reduced the diversity and changed the community composition of protists. Therefore, protist communities turned out as the most susceptible microbiome group to changes in soil conditions as induced by fertilization. Nevertheless, the changes in bacterial and fungal communities were found to be important bottom-up forces driving protist diversity. Several protist taxa were identified as potential keystone microbiome taxa, which might control microbiome functioning through top-down forces. This finding is of far-reaching importance to understand microbiome communities and consequently system’s stability. Future microbiome and soil biodiversity studies in general should include protist community analyses as otherwise important information on microbiome structures might be missing. Overall, we propose that a reduction of protist diversity induced by application of nitrogen fertilizers, especially if applied intensively, might have detrimental effects on agricultural soil ecosystem functioning and sustainability, especially over the long term.
Field site setup, management, and samplings
Three field experimental sites were located across a latitudinal gradient in China from north to central to south. The sites were chosen to represent the three main agricultural production areas in China (Additional file 1: Figure S7) and consisted of black soil in Gongzhuling (Jilin Province 43° 29′ N, 124° 47′ E, Mollisols), fluvo-aquic soil in Xuchang (Henan Province, 34° 2′ N, 113° 51′ E, Calcaric cambisol), and red soil in Taoyuan (Hunan Province, 28° 54′ N, 111° 29′ E, Ultisol). The average annual temperature in Gongzhuling (GZL), Xuchang (XC), and Taoyuan (TY) site was 6.1 °C, 14.4 °C, and 17.4 °C, respectively, and the wheat-maize rotation system was applied in Xuchang (XC) and Taoyuan (TY), while only maize was planted in Gongzhuling (GZL) due to the harsh winter. The geographical map (Additional file 1: Figure S7) was generated by the “maps,” “maptools,” “grid,” “ggsn,” “legendMap,” and “ggplot” packages in R platform. Three treatments, including no nitrogen fertilizer addition (control), nitrogen addition (N), and nitrogen plus straw addition (NS), were set up in triplicate plots (3.125 m × 8 m each plot) in each field site since 2015. All treatments were amended with the same dose of phosphorus (120 kg ha−1 P2O5) and potassium (120 kg ha−1 K2O) fertilizers. Nitrogen fertilizers were applied with an annual rate of 200 kg ha−1 nitrogen, and straw was applied with a rate of 5000 kg ha−1, consisting of dry maize straw from the last harvest that was cut into 2–3-cm lengths. All fertilizers and straw were used as a basal fertilizer before sowing and were hand applied in furrows and then covered as ridges.
Sampling was carried out at the maize heading stage in summer and before harvest in autumn after 2 years fertilization in 2016. Each sample was a composite of five cores randomly collected from the ridge. In each core, soil samples were collected from 5- to 10-cm depth on the ridge and between two plants, with top 5-cm soil layer removed to avoid exogenous disturbance (Additional file 1: Figure S7). In total, 54 soil samples were collected and immediately carried back to the laboratory on ice. Raw soil samples were sieved through 2-mm-diameter mesh. In addition, fine root and plant residue that can pass through the mesh was avoided before laboratory analysis.
Soil physicochemical properties
Soil pH was determined in a 1:2.5 soil/water suspension with a glass electrode. Soil moisture was measured by the loss of weight after oven drying at 105 °C to constant weight. Total C (TC) and N (TN) were analyzed by an elemental analyzer (Vario EL III-Elementar, Germany). C/N was calculated by the ratio of TC and TN. Organic matter (OM) was measured using the K2Cr2O7-H2SO4 oxidation-reduction colorimetric method . Dissolved organic carbon (DOC) was extracted by 0.5 M K2SO4 and measured by a TOC analyzer (Multi N/C 3100, Analytikjena, German). NO3−-N and NH4+-N were extracted with 1 M KCl and measured using a continuous flow analyzer (AA3, SEAL analytical, Germany).
DNA extraction, PCR assays, and high-throughput sequencing
Soil subsamples for molecular analysis were kept in a freezer at − 80 °C before use. Total DNA was extracted from 0.4 g of soil using the PowerSoil DNA Isolation Kit (MO BIO laboratories, Carlsbad, USA). DNA quantity and quality were determined using a NanoDrop Spectrophotometer (NanoDrop Technologies Inc., Wilmington, DE, USA). Three commonly used primer sets were applied to metabarcoding approaches to study microbial communities targeting bacterial 16S rRNA genes , fungal  and protists  18S rRNA genes, respectively (Additional file 1: Table S1). We acknowledge that the ITS region provides higher taxonomic resolution of fungal community , but we chose the 18S rRNA gene to allow phylogenetic tree construction  to acquire phylogenetic diversity of microbiomes . PCR reactions were conducted in four parallels in 25 μl mixtures consisting of 12.5 μl Premix Taq™ (Takara Biotechnology, Dalian, China), 0.5 μl of each primer (10 μM), 2 μl of template which is five times diluted of extracted DNA, and 9.5 μl of sterilized ddH2O. Negative control samples were also included throughout the PCR assay to ensure reaction systems were not contaminated. PCR conditions for each primer set are detailed in Additional file 1: Table S1. PCR amplicons were extracted from 2% agarose gels and purified by using of the AxyPrep DNA Gel Extraction Kit (Axygen Biosciences, Union City, CA, USA) according to the manufacturer’s instructions and quantified through QuantiFluor™ -ST (Promega, USA). Purified amplicons were pooled in equimolar and sent for paired-end sequencing on an Illumina MiSeq PE 300 × 2 sequencer (Majorbio Bio-Pharm Technology Co. Ltd., Shanghai, China).
Quantitative Insights into Microbial Ecology (QIIME) 1.90  standard operation procedure was used to process raw sequences. In short, raw reads of individual samples were merged to paired-end reads (multiple_join_paired_ends.py), followed by removing barcodes from sequences (multiple_extract_barcodes.py) and demultiplex and quality filter sequence from data files (multiple_split_libraries_fastq.py). UPARSE was used for chimera removal, and operational taxonomic units (OTUs) were clustered at 97% sequence similarity and a representative sequence of each OTU was selected and used for taxonomic assignments . Protist OTUs were taxonomically assigned by blasting against the Protist Ribosomal Reference (PR2) database (version_4.5) , while bacterial and fungal OTUs were assigned against the SILVA database (version_123)  at 90% minimum similarity (assign_taxonomy.py). Sequences obtained were rarefied at minimum number of sequences per sample (bacteria, 41,359; fungi, 37,002; eukaryotes, 36,242) for downstream analysis (single_rarefaction.py). Rhodophyta, Streptophyta, Fungi, Opisthokonta_X, Metazoa, and ambiguous taxa in Eukaryotes were excluded from the obtained protist OTU table (filter_taxa_from_otu_table.py). Representative sequences were aligned against the SILVA database  (align_seqs.py) to acquire the phylogenetic tree (make_phylogeny.py). Alpha diversity (faith_pd index) of bacterial, fungal, and protist communities were calculated based on the faith’s phylogenetic metric at OTU level (alpha_diversity.py). Nonmetric multidimensional scaling (NMDS) (nmds.py) was used to visualize dissimilarity of beta diversity based on the unweighted unifrac distance across different treatments according to the phylogenetic tree (beta_diversity.py) .
In order to determine the effects of the application of nitrogen fertilizers on soil microbiome associations in the three types of soils, the underlying co-occurrences among protist, bacterial, and fungal taxa was depicted through network analysis using the CoNet plug-in in Cytoscape [64, 65]. The network analysis was performed at the family level of these three microbiomes to reduce the complexity of calculation and to ensure the accuracy of taxonomic information. Data filtering was performed prior to network construction to avoid zero values that could result in spurious correlations and to mitigate the seasonal variances, and therefore, the taxa represented in all samples were reserved . To explore all the pairwise associations, correlation scores (Spearman correlation, Pearson correlation, Kullback-Leibler dissimilarity, Bray-Curtis dissimilarity and mutual information) were calculated [67, 68]. To avoid potential false-positive correlations and compositionality biases, the ReBoot procedure with 100 permutations was performed, and the resultant distribution was refined with 1000 bootstraps . The P values of the five methods were integrated by the Brown method, and only significant correlations (P < 0.05) were retained for the downstream procedure . The Benjamini-Hochberg multiple test was performed as a correlation after the Brown merging P values, which adjusted the false discovery rate, and the chance of false rejecting the null hypothesis is ≤ 0.05 . The resulting correlations were imported into the Gephi platform and then visualized by the Frucherman Reingold algorithms . The topology property parameters of the network, the degree, betweenness centrality, and closeness centrality of each node in the network were calculated by the plug-in Network Analyzer in Cytoscape . The clustering coefficient and network density were chosen to reflect the changes of soil microbiomes associations as they are measurements of how close the nodes are embedded in their neighborhood and clustered together . Degree, betweenness centrality (BC), and closeness centrality (CC) were selected to explore key hubs of networks . The node with higher degree indicates it is highly connected with other nodes, and the node with higher BC value indicates that it is more closely connected to nodes in other modules of the network and is a candidate for a connector of the network. The node with higher CC value indicates it is more closely connected within the module of the network and is a candidate for a module hub of the network, and the node with higher both BC and CC values indicates it is the super-generalist of the network and a candidate for a hub of the network [74, 75]. Here, the cutoff value of the BC and CC of a node was set at 0.6 to explore the putative key hubs of each network. Furthermore, in order to investigate associations between protists and other microorganisms that can reflect the potential predatory-prey interactions, the edges linking protist to the bacterial or fungal taxa in at least two networks were calculated.
Network analysis was also used to determine the difference in soil microbiome associations between summer and autumn. Detailed descriptions are provided as supplementary information.
Structural equation models
In order to quantify the importance of nitrogen addition and nitrogen plus straw addition on the changes of soil physicochemical properties and alpha diversity of bacterial, fungal, and protist communities in soils, SEMs were constructed. Based on biogeographical knowledge at present, the a priori and theoretical model assumed that (i) nitrogen fertilizers (N and NS treatments) directly influence alpha diversity of soil bacterial, fungal, and protist communities, respectively; (ii) nitrogen fertilizers indirectly affect alpha diversity of soil bacterial, fungal, and protist communities by changing soil physicochemical properties; and (iii) nitrogen fertilizers indirectly influence alpha diversity of soil bacterial, fungal, and protist communities by altering inter-kingdom interactions within soil microbiomes as a result of competition and predatory-prey interactions between the different microbial groups. Nitrogen addition (N) variables were created by assigning the value 1 to the nitrogen addition treatment and 0 to the control and nitrogen plus straw addition treatments. The nitrogen plus straw addition (NS) variables were created by assigning the value 1 to the nitrogen plus straw addition treatment and 0 to the control and nitrogen addition treatments. All of the measured soil physicochemical property indices were reduced in dimensions by nonmetric multidimensional scaling (NMDS), with the variances of soil properties being represented on the first axis of NMDS. Faith’s phylogenetic diversity metric was used to determine the alpha diversity of bacterial, fungal, and protist community at OUT level. All variables were standardized by Z transformation (mean = 0, standard deviation = 1) to improve normality using the scale function in R. The pairwise correlation among these variables was calculated by the Mantel test using the “Ecodist” package in R platform, and a covariance matrix of these variables was inserted into AMOS 17.0 (SPSS, Chicago, IL, USA) for SEM construction and analysis. Maximum likelihood estimation was used to fit the covariance matrix to the model . Chi-square (P > 0.05), goodness-of-fit-index (GFI > 0.90), and root mean square error of approximation (RMSEA < 0.05) were measured to ensure the model adequately fit .
SEMs were also used to quantify effects of abiotic factors (nitrogen fertilizers, soil physicochemical properties) and biotic factors (bacterial community, fungal community) on the significant reductions (P < 0.05) in alpha diversity of protist community in the black soil in summer, fluvo-aquic soil in autumn, and red soil in summer and autumn. Detailed descriptions are provided as supplementary information.
Statistical analyses were performed in R platform (3.3.1). One-way ANOVA was used to analyze differences in soil physicochemical properties, and alpha diversity among nitrogen fertilizer treatments and plant growth seasons (packages: agricolae, car), statistical differences were considered significant at P < 0.05. Duncan post hoc test was used to assess treatment differences in one-way ANOVA analyses. Variables were standardized by Z transformation (mean = 0, standard deviation = 1) to improve normality using scale function in R. Linear mixed models (LMM) were used to analyze the effects of nitrogen fertilization, soil type, season, and their interactive effects on alpha diversity of soil microbiome, with the plot position serial number (i _ j, where i and j are the row and column number of the plot, respectively) in the field as a random effect (packages: lme4, lmerTest) . Permutation multivariate analysis of variance (PERMANOVA) was employed to assess the significance of the influential factors that differentiate soil physicochemical properties and microbiomes (packages: vegan, adonis function) . The soil physicochemical properties data was input as column list of each variable (pH, moisture, etc.) for PERMANOVA followed by Bray-Curtis distance calculation, and the microbiomes data was input as matrix based on unweighted unifrac distance. Linear discriminant analysis effect size (LEfSe) was performed to investigate potential biomarkers (across five taxonomic levels, from phylum to genus for bacterial and fungal communities, from supergroup to family level for protist community) within soil microbiomes specifically enrich in one of the treatments in each soil type, in summer and autumn, respectively, based on P < 0.05 and a LDA score > 2.0 . Biomarkers were color labeled on cladograms according to the PR2 taxonomy. Spearman correlations were calculated for depicting the relationship between the alpha diversity of the microbiomes and physicochemical properties in soils.
We thank Prof. Yanlai Han, Drs. Zhi Quan, Rong Sheng and Shanlong Li for cooperation and assistance in the field experimentation, Chuanfa Wu and Pengxia Xu for laboratory work, Shuai Du, Bing Han and Chao Xiong for their assist in data processes.
This work was financially supported by the Strategic Priority Research Program of the Chinese Academy of Sciences (XDB15020200) and the National Key R&D Program (2017YFD0200600). Stefan Geisen was supported by a NWO-VENI grant from the Netherlands Organisation for Scientific Research (016.Veni.181.078), and Li-Mei Zhang by the Youth Innovation Promotion Association, Chinese Academy of Sciences.
Availability of data and materials
The sequencing reads of the bacterial 16S rRNA gene, fungal 18S rRNA gene and protist 18S rRNA gene were deposited in the Sequence Read Archive of NCBI under the accession number PRJNA498197, PRJNA498206 and PRJNA498211, respectively.
JZH, LMZ, WXW, YTF, and ZBZ designed the research. LMZ, LLH, JPS, WXW, YTF, and PPL managed the field trial stations. ZBZ sampled the soils and conducted the laboratory analyses and the raw data collection. ZBZ, SG, JTW, and LMZ performed the data processes. ZBZ, SG, LMZ, and JZH wrote the manuscript. All authors read and approved the final manuscript.
Ethics approval and consent to participate
Not applicable as there were no human, animal or pathogen subjects involved.
Consent for publication
The authors declare that they have no competing interests.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
- Lal R. Soil carbon sequestration impacts on global climate change and food security. Science. 2004;304:1623–7.View ArticleGoogle Scholar
- Wagg C, Bender SF, Widmer F, van der Heijden MG. Soil biodiversity and soil community composition determine ecosystem multifunctionality. Proc Natl Acad Sci U S A. 2014;111:5266–70.View ArticleGoogle Scholar
- Loreau M. Biodiversity and ecosystem functioning: a mechanistic model. Proc Natl Acad Sci U S A. 1998;95:5632–6.View ArticleGoogle Scholar
- Hu HW, He JZ. Manipulating the soil microbiome for improved nitrogen management. Microbiology Australia. 2018;39:24–7.Google Scholar
- Philippot L, Raaijmakers JM, Lemanceau P, van der Putten WH. Going back to the roots: the microbial ecology of the rhizosphere. Nat Rev Microbiol. 2013;11:789–99.View ArticleGoogle Scholar
- Bender SF, Wagg C, van der Heijden MGA. An underground revolution: biodiversity and soil ecological engineering for agricultural sustainability. Trends Ecol Evol. 2016;31:440–52.View ArticleGoogle Scholar
- Berendsen RL, Pieterse CM, Bakker PA. The rhizosphere microbiome and plant health. Trends Plant Sci. 2012;17:478–86.View ArticleGoogle Scholar
- McLaughlin A, Mineau P. The impact of agricultural practices on biodiversity. Agric Ecosyst Environ. 1995;55:201–12.View ArticleGoogle Scholar
- Sun R, Zhang XX, Guo X, Wang D, Chu H. Bacterial diversity in soils subjected to long-term chemical fertilization can be more stably maintained with the addition of livestock manure than wheat straw. Soil Biol Biochem. 2015;88:9–18.View ArticleGoogle Scholar
- Zhang LM, Offre PR, He JZ, Verhamme DT, Nicol GW, Prosser JI. Autotrophic ammonia oxidation by soil thaumarchaea. Proc Natl Acad Sci U S A. 2010;107:17240–5.View ArticleGoogle Scholar
- Hu H-W, Macdonald CA, Trivedi P, Anderson IC, Zheng Y, Holmes B, et al. Effects of climate warming and elevated CO2 on autotrophic nitrification and nitrifiers in dryland ecosystems. Soil Biol Biochem. 2016;92:1–15.View ArticleGoogle Scholar
- Payne RJ, Thompson AM, Standen V, Field CD, Caporn SJM. Impact of simulated nitrogen pollution on heathland microfauna, mesofauna and plants. Eur J Soil Biol. 2012;49:73–9.View ArticleGoogle Scholar
- Wang C, Liu D, Bai E. Decreasing soil microbial diversity is associated with decreasing microbial biomass under nitrogen addition. Soil Biol Biochem. 2018;120:126–33.View ArticleGoogle Scholar
- Heemsbergen DA, Berg MP, Loreau M, van Hal JR, Faber JH, Verhoef HA. Biodiversity effects on soil processes explained by interspecific functional dissimilarity. Science. 2004;306:1019–20.View ArticleGoogle Scholar
- Philippot L, Spor A, Henault C, Bru D, Bizouard F, Jones CM, et al. Loss in microbial diversity affects nitrogen cycling in soil. ISME J. 2013;7:1609–19.View ArticleGoogle Scholar
- Quan Z, Li S, Zhu F, Zhang L, He J, Wei W, et al. Fates of 15N-labeled fertilizer in a black soil-maize system and the response to straw incorporation in Northeast China. J Soils Sediments. 2017;18:1441–52.View ArticleGoogle Scholar
- Birkhofer K, Bezemer TM, Bloem J, Bonkowski M, Christensen S, Dubois D, et al. Long-term organic farming fosters below and aboveground biota: implications for soil quality, biological control and productivity. Soil Biol Biochem. 2008;40:2297–308.View ArticleGoogle Scholar
- Zeng J, Liu X, Song L, Lin X, Zhang H, Shen C, et al. Nitrogen fertilization directly affects soil bacterial diversity and indirectly affects bacterial community composition. Soil Biol Biochem. 2016;92:41–9.View ArticleGoogle Scholar
- Geisen S, Mitchell EAD, Adl S, Bonkowski M, Dunthorn M, Ekelund F, et al. Soil protists: a fertile frontier in soil biology research. FEMS Microbiol Rev. 2018;42:293–323.View ArticleGoogle Scholar
- Krashevska V, Sandmann D, Maraun M, Scheu S. Moderate changes in nutrient input alter tropical microbial and protist communities and belowground linkages. ISME J. 2014;8:1126–34.View ArticleGoogle Scholar
- Scharroba A, Dibbern D, Hünninghaus M, Kramer S, Moll J, Butenschoen O, et al. Effects of resource availability and quality on the structure of the micro-food web of an arable soil across depth. Soil Biol Biochem. 2012;50:1–11.View ArticleGoogle Scholar
- Guo S, Xiong W, Xu H, Hang X, Liu H, Xun W, et al. Continuous application of different fertilizers induces distinct bulk and rhizosphere soil protist communities. Eur J Soil Biol. 2018;88:8–14.View ArticleGoogle Scholar
- Lentendu G, Wubet T, Chatzinotas A, Wilhelm C, Buscot F, Schlegel M. Effects of long-term differential fertilization on eukaryotic microbial communities in an arable soil: a multiple barcoding approach. Mol Ecol. 2014;23:3341–55.View ArticleGoogle Scholar
- Mieczan T. Distributions of testate amoebae and ciliates in different types of peatlands and their contributions to the nutrient supply. Zool Stud. 2012;51:18–26.Google Scholar
- Bates ST, Clemente JC, Flores GE, Walters WA, Parfrey LW, Knight R, et al. Global biogeography of highly diverse protistan communities in soil. ISME J. 2013;7:652–9.View ArticleGoogle Scholar
- Geisen S, Cornelia B, Jörg R, Michael B. Soil water availability strongly alters the community composition of soil protists. Pedobiologia. 2014;57:205–13.View ArticleGoogle Scholar
- Grossmann L, Jensen M, Heider D, Jost S, Glucksman E, Hartikainen H, et al. Protistan community analysis: key findings of a large-scale molecular sampling. ISME J. 2016;10:2269–79.View ArticleGoogle Scholar
- Zahn G, Wagai R, Yonemura S. The effects of amoebal bacterivory on carbon and nitrogen dynamics depend on temperature and soil structure interactions. Soil Biol Biochem. 2016;94:133–7.View ArticleGoogle Scholar
- Ju XT, Xing GX, Chen XP, Zhang SL, Zhang LJ, Liu XJ, et al. Reducing environmental risk by improving N management in intensive Chinese agricultural systems. Proc Natl Acad Sci U S A. 2009;106:3041–6.View ArticleGoogle Scholar
- Schulz-Bohm K, Geisen S, Wubs ER, Song C, de Boer W, Garbeva P. The prey’s scent - volatile organic compound mediated interactions between soil bacteria and their protist predators. ISME J. 2017;11:817–20.View ArticleGoogle Scholar
- Bonkowski M. Protozoa and plant growth: the microbial loop in soil revisited. New Phytol. 2004;162:617–31.View ArticleGoogle Scholar
- de Ruiter PC, Neutel AM, Moore JC. Energetics, patterns of interaction strengths, and stability in real ecosystems. Science. 1995;269:1257–60.View ArticleGoogle Scholar
- Payne RJ. Seven reasons why protists make useful bioindicators. Acta Protozool. 2013;52:105–13.Google Scholar
- Zhou J, Jiang X, Zhou B, Zhao B, Ma M, Guan D, et al. Thirty four years of nitrogen fertilization decreases fungal diversity and alters fungal community composition in black soil in northeast China. Soil Biol Biochem. 2016;95:135–43.View ArticleGoogle Scholar
- Zhou Z, Wang C, Zheng M, Jiang L, Luo Y. Patterns and mechanisms of responses by soil microbial communities to nitrogen addition. Soil Biol Biochem. 2017;115:433–41.View ArticleGoogle Scholar
- Fierer N. Embracing the unknown: disentangling the complexities of the soil microbiome. Nat Rev Microbiol. 2017;15:579–90.View ArticleGoogle Scholar
- Eisenhauer N, Bessler H, Engels C, Gleixner G, Habekost M, Milcu A, et al. Plant diversity effects on soil microorganisms support the singular hypothesis. Ecology. 2010;91:485–96.View ArticleGoogle Scholar
- Delgado-Baquerizo M, Oliverio AM, Brewer TE, Benavent-Gonzalez A, Eldridge DJ, Bardgett RD, et al. A global atlas of the dominant bacteria found in soil. Science. 2018;359:320.View ArticleGoogle Scholar
- Wu W, Lu HP, Sastri A, Yeh YC, Gong GC, Chou WC, et al. Contrasting the relative importance of species sorting and dispersal limitation in shaping marine bacterial versus protist communities. ISME J. 2018;12:485–94.View ArticleGoogle Scholar
- Murase J, Hida A, Ogawa K, Nonoyama T, Yoshikawa N, Imai K. Impact of long-term fertilizer treatment on the microeukaryotic community structure of a rice field soil. Soil Biol Biochem. 2015;80:237–43.View ArticleGoogle Scholar
- Lu C, Chen H, Teng Z, Yuan L, Ma J, He H, et al. Effects of N fertilization and maize straw on the dynamics of soil organic N and amino acid N derived from fertilizer N as indicated by 15N labeling. Geoderma. 2018;321:118–26.View ArticleGoogle Scholar
- Lu C, Wang H, Chen H, Yuan L, Ma J, Shi Y, et al. Effects of N fertilization and maize straw on the transformation and fate of labeled ((NH4)-15N)(2) SO4 among three continuous crop cultivations. Agr Water Manage. 2018;208:275–83.View ArticleGoogle Scholar
- Eisenhauer N, Cesarz S, Koller R, Worm K, Reich PB. Global change belowground: impacts of elevated CO2, nitrogen, and summer drought on soil food webs and biodiversity. Glob Chang Biol. 2012;18:435–47.View ArticleGoogle Scholar
- Friman VP, Dupont A, Bass D, Murrell DJ, Bell T. Relative importance of evolutionary dynamics depends on the composition of microbial predator-prey community. ISME J. 2016;10:1352–62.View ArticleGoogle Scholar
- Saleem M, Fetzer I, Dormann CF, Harms H, Chatzinotas A. Predator richness increases the effect of prey diversity on prey yield. Nat Commun. 2012;3:1305.View ArticleGoogle Scholar
- Geisen S, Koller R, Hünninghaus M, Dumack K, Urich T, Bonkowski M. The soil food web revisited: diverse and widespread mycophagous soil protists. Soil Biol Biochem. 2016;94:10–8.View ArticleGoogle Scholar
- Banerjee S, Schlaeppi K, van der Heijden MGA. Keystone taxa as drivers of microbiome structure and functioning. Nat Rev Microbiol. 2018;16:567–76.View ArticleGoogle Scholar
- Wang S, Wang X, Han X, Deng Y. Higher precipitation strengthens the microbial interactions in semi-arid grassland soils. Glob Ecol Biogeogr. 2018;27:570–80.View ArticleGoogle Scholar
- Shi S, Nuccio EE, Shi ZJ, He Z, Zhou J, Firestone MK. The interconnected rhizosphere: high network complexity dominates rhizosphere assemblages. Ecol Lett. 2016;19:926–36.View ArticleGoogle Scholar
- Yu J, Deem LM, Crow SE, Deenik JL, Penton CR. Biochar application influences microbial assemblage complexity and composition due to soil and bioenergy crop type interactions. Soil Biol Biochem. 2018;117:97–107.View ArticleGoogle Scholar
- Morrien E, Hannula SE, Snoek LB, Helmsing NR, Zweers H, de Hollander M, et al. Soil networks become more connected and take up more carbon as nature restoration progresses. Nat Commun. 2017;8:14349.View ArticleGoogle Scholar
- Li J, Ma YB, Hu HW, Wang JT, Liu YR, He JZ. Field-based evidence for consistent responses of bacterial communities to copper contamination in two contrasting agricultural soils. Front Microbiol. 2015;6:31.PubMedPubMed CentralGoogle Scholar
- Bates ST, Berg-Lyons D, Caporaso JG, Walters WA, Knight R, Fierer N. Examining the global distribution of dominant archaeal populations in soil. ISME J. 2011;5:908–17.View ArticleGoogle Scholar
- Rousk J, Baath E, Brookes PC, Lauber CL, Lozupone C, Caporaso JG, et al. Soil bacterial and fungal communities across a pH gradient in an arable soil. ISME J. 2010;4:1340–51.View ArticleGoogle Scholar
- Stoeck T, Bass D, Nebel M, Christen R, Jones MD, Breiner HW, et al. Multiple marker parallel tag environmental DNA sequencing reveals a highly complex eukaryotic community in marine anoxic water. Mol Ecol. 2010;19(Suppl 1):21–31.View ArticleGoogle Scholar
- Schoch CL, Seifert KA, Huhndorf S, Robert V, Spouge JL, Levesque CA, et al. Nuclear ribosomal internal transcribed spacer (ITS) region as a universal DNA barcode marker for Fungi. Proc Natl Acad Sci U S A. 2012;109:6241–6.View ArticleGoogle Scholar
- Price MN, Dehal PS, Arkin AP. FastTree: computing large minimum evolution trees with profiles instead of a distance matrix. Mol Biol Evol. 2009;26:1641–50.View ArticleGoogle Scholar
- Washburne AD, Morton JT, Sanders J, McDonald D, Zhu Q, Oliverio AM, et al. Methods for phylogenetic analysis of microbiome data. Nat Microbiol. 2018;3:652–61.View ArticleGoogle Scholar
- Caporaso JG, Kuczynski J, Stombaugh J, Bittinger K, Bushman FD, Costello EK, et al. QIIME allows analysis of high-throughput community sequencing data. Nat Methods. 2010;7:335–6.View ArticleGoogle Scholar
- Edgar RC. UPARSE: highly accurate OTU sequences from microbial amplicon reads. Nat Methods. 2013;10:996–8.View ArticleGoogle Scholar
- Guillou L, Bachar D, Audic S, Bass D, Berney C, Bittner L, et al. The Protist Ribosomal Reference database (PR2): a catalog of unicellular eukaryote small sub-unit rRNA sequences with curated taxonomy. Nucleic Acids Res. 2013;41:D597–604.View ArticleGoogle Scholar
- Quast C, Pruesse E, Yilmaz P, Gerken J, Schweer T, Yarza P, et al. The SILVA ribosomal RNA gene database project: improved data processing and web-based tools. Nucleic Acids Res. 2013;41:D590–6.View ArticleGoogle Scholar
- Kenkel NC, Orloci L. Applying metric and nonmetric multidimensional scaling to ecological studies: some new results. Ecology. 1986;67:919–28.View ArticleGoogle Scholar
- Soffer N, Zaneveld J, Vega TR. Phage-bacteria network analysis and its implication for the understanding of coral disease. Environ Microbiol. 2015;17:1203–18.View ArticleGoogle Scholar
- Shannon P, Markiel A, Ozier O, Baliga NS, Wang JT, Ramage D, et al. Cytoscape: a software environment for integrated models of biomolecular interaction networks. Genome Res. 2003;13:2498–504.View ArticleGoogle Scholar
- Hu HW, Wang JT, Li J, Li JJ, Ma YB, Chen D, et al. Field-based evidence for copper contamination induced changes of antibiotic resistance in agricultural soils. Environ Microbiol. 2016;18:3896–909.View ArticleGoogle Scholar
- Faust K, Sathirapongsasuti JF, Izard J, Segata N, Gevers D, Raes J, et al. Microbial co-occurrence relationships in the human microbiome. PLoS Comput Biol. 2012;8:e1002606.View ArticleGoogle Scholar
- Faust K, Raes J. Microbial interactions: from networks to models. Nat Rev Microbiol. 2012;10:538–50.View ArticleGoogle Scholar
- Hu HW, Wang JT, Li J, Shi XZ, Ma YB, Chen D, et al. Long-term nickel contamination increases the occurrence of antibiotic resistance genes in agricultural soils. Environ Sci Technol. 2017;51:790–800.View ArticleGoogle Scholar
- Jacomy M, Bastian M, Heymann S. Gephi: an open source software for exploring and manipulating networks. Int AAAI Conf Weblogs Soc Media. 2009.Google Scholar
- Assenov Y, Ramirez F, Schelhorn SE, Lengauer T, Albrecht M. Computing topological parameters of biological networks. Bioinformatics. 2008;24:282–4.View ArticleGoogle Scholar
- Barberan A, Bates ST, Casamayor EO, Fierer N. Using network analysis to explore co-occurrence patterns in soil microbial communities. ISME J. 2012;6:343–51.View ArticleGoogle Scholar
- Sporns O, Honey CJ, Kotter R. Identification and classification of hubs in brain networks. PLoS One. 2007;2:e1049.View ArticleGoogle Scholar
- Martin Gonzalez AM, Dalsgaard B, Olesen JM. Centrality measures and the importance of generalist species in pollination networks. Ecol Complex. 2010;7:36–43.View ArticleGoogle Scholar
- Deng Y, Jiang YH, Yang Y, He Z, Luo F, Zhou J. Molecular ecological network analyses. BMC Bioinformatics. 2012;13:113.View ArticleGoogle Scholar
- Jun-Tao Wang, Yuan-Ming Zheng, Hang-Wei Hu, Jing Li, Li-Mei Zhang, Bao-Dong Chen, Wei-Ping Chen, Ji-Zheng He. Coupling of soil prokaryotic diversity and plant diversity across latitudinal forest ecosystems. Scientific Reports 2016;6:1.Google Scholar
- Grace JB, Keeley JE. A structural equation model analysis of postfire plant diversity in California shrublands. Ecol Appl. 2006;16:503–14.View ArticleGoogle Scholar
- Wagner MR, Lundberg DS, Del Rio TG, Tringe SG, Dangl JL, Mitchell-Olds T. Host genotype and age shape the leaf and root microbiomes of a wild perennial plant. Nat Commun. 2016;7:12151.View ArticleGoogle Scholar
- Bell TH, Hassan SE, Lauron-Moreau A, Al-Otaibi F, Hijri M, Yergeau E, et al. Linkage between bacterial and fungal rhizosphere communities in hydrocarbon-contaminated soils is related to plant phylogeny. ISME J. 2014;8:331–43.View ArticleGoogle Scholar
- Segata N, Izard J, Waldron L, Gevers D, Miropolsky L, Garrett WS, et al. Metagenomic biomarker discovery and explanation. Genome Biol. 2011;12:R60.View ArticleGoogle Scholar