Composition and diversity of bacterial communities in the rhizosphere of the Chinese medicinal herb Dendrobium

Background Dendrobium is a precious herbal that belongs to Orchidaceae and is widely used as health care traditional Chinese medicine in Asia. Although orchids are mycorrhizal plants, most research still focuses on endophytes, and there is still large amount unknown about rhizosphere microorganisms. To investigate the rhizosphere microbial community of different Dendrobium species during the maturity stage, we used high-throughput sequencing to analyze microbial community in rhizosphere soil during the maturity stage of three kinds of Dendrobium species. Results In our study, a total of 240,320 sequences and 11,179 OTUs were obtained from these three Dendrobium species. According to the analysis of OTU annotation results, different Dendrobium rhizosphere soil bacteria include 2 kingdoms, 63 phyla, 72 classes, 159 orders, 309 families, 850 genera and 663 species. Among all sequences, the dominant bacterial phyla (relative abundance > 1%) were Proteobacteria, Actinobacteria, Bacteroidetes, Acidobacteria, Firmicutes, Verrucomicrobia, Planctomycetes, Chloroflexi, and Gemmatimonadetes. And through WGCNA analysis, we found the hub flora was also belong to Acidobacteria, Actinobacteria and Proteobacteria. Conclusions We found that the rhizosphere bacterial communities of the three kinds of Dendrobium have significant differences, and that the main species of rhizosphere microorganisms of Dendrobium are concentrated in the Proteobacteria, Actinobacteria, and Bacteroidetes. Moreover, the smaller the bacterial level, the greater the difference among Dendrobium species. These results fill knowledge gaps in the rhizosphere microbial community of Dendrobium and provide a theoretical basis for the subsequent mining of microbial functions and the study of biological fertilizers. Supplementary Information The online version contains supplementary material available at 10.1186/s12870-021-02893-y.


Background
Rhizosphere soil refers to the narrow zone of soil affected by root exudations, containing up to 10 11 microbial cells and over 30,000 prokaryotic species [1][2][3]. Rhizosphere microbes are significantly different from non-rhizosphere microbes in terms of species, number, and activity [4]. There are abundant microbial resources in rhizosphere soil, which can be 10-1000 times that in bulk soil [5]. The rhizosphere bacterial communities are represented by diverse bacterial taxa, though their abundance varies across the root system under different soil and plant types [6]. In addition, the contents of the Firmicutes and Acidobacteria are also high [7]. Using high-throughput sequencing, Lundberg et al. found that Arabidopsis thaliana had the highest relative abundance of Proteobacteria in the rhizosphere bacterial community, with Pseudomonadaceae being the main population, followed by Bacteroides and Actinomyces [8]. The interaction between rhizosphere microbes and plants can affect material circulation and energy flow. Changes in its community structure and abundance can affect plant growth and development, flowering and fruiting, and plant interaction with phytophagous insects, which is of great significance for plant growth and yield [9,10]. Dendrobium spp. is the second largest genus in Orchidaceae, with more than 1400 species in the world. There are 74 species and 2 varieties of Dendrobium in Orchidaceae in China, most of which are precious medicinal plants. Dendrobium has many functions, such as benefiting stomach and promoting body fluid, clearing heat and nourishing yin, relieving inflammation and relieving pain, clearing eyesight, and enhancing immunity [11]. Currently, artificial cultivation using non-symbiotic tissue culture does not meet market demand because of slow growth and low survival rates [12]. Therefore, the development of an effective method for propagating these endangered species for both conservation and commercial production is needed. There are three kinds of medicinal Dendrobium in the Ta-pieh Mountain area, which are Dendrobium huoshanense, Dendrobium officinale and Dendrobium moniliforme [13]. All of them are perennial herbs of Dendrobium in Orchidaceae. They contain mainly polysaccharides, alkaloids and bibenzyls. They have antitumor, immunomodulatory, antioxidant, vasodilator and hypoglycemic effects [14].
At present, the research in Dendrobium mainly focuses on the endophytic bacteria. Indeed, there is a complex relationship between Dendrobium and its endophytes. Previous studies have shown that Sphiugomouas and Mycobacterium bacteria isolated from the roots of D. moschatum (Buch. -ham) SW. could significantly improve the seed germination rate of D. moschatum (Buch. -ham) SW [15]; and there are many studies using protocorm as material to successfully isolate and obtain effective fungi that promote seed germination [16][17][18]. However, few reports are available regarding Dendrobium and soil microbial communities. There are abundant bacteria, fungi and actinomycetes in the rhizosphere of medicinal plants. Not only can they promote the absorption of soil nutrients by medicinal plants, but they can also improve the yield and quality of medicinal plants; however, they can also cause continuous cropping obstacles for medicinal plants. So, studying Dendrobium medicinal plants and their rhizosphere microbes is of great significance to clarify how rhizosphere microbes can improve the yield and quality of medicinal plants. However, the traditional cultivation methods account for 0.1 to 1% of the environmental microorganisms, which cannot fully reflect the real situation of the environmental microbial community. In recent years, with the rapid development of new generation sequencing technology, based on 16 s rRNA sequence amplification and Illumina Miseq high-throughput sequencing technology, huge amounts of data can be obtained. By means of bioinformatics, the overall microbial community composition can be obtained. With the advantages of high-throughput, low price and short operation cycle, it has been widely used in the study of microbial community structure. Especially in the field of medicinal plants, such as Panax ginseng [19], Panax notoginseng [20], Ajuga bracteosa [21], Origanum vulgare [22], Lilium davidii [23] and other plants,these methods are widely used.
They always interact with other microorganisms such as mutualism, parasitism, antagonism. Co-culture is often used to study the interaction between microorganisms [24], but due to the inculcability of most microorganisms, co-culture has shown great limitations. Therefore, it is necessary to find other reliable ways to study the interaction between microbial communities. The rapid development of high-throughput sequencing has provided a new method for the study of microbial interaction. Based on high-throughput sequencing results, researchers have developed a number of software to predict the interaction among microbial communities, such as CoNet, MENA [25]. These network analyses reflect the co-expression relationship among microbial communities. Weighted gene co-expression network analysis (WGCNA) is a kind of gene expression can be measured with a specific character combined with the analysis of network analysis algorithm [26]. It can divide the gene co-expression network of complex biological processes into several highly correlated characteristic modules, and can establish the correlation between the modules and specific traits, so as to find the hub flora. WGCNA analysis has been widely used in transcriptional data analysis of various plants (such as corn [27], tomato [28] and Dioscorea nipponica [29]). But now, WGCNA analysis also has many applications in microbiome, such as rhizosphere microbes of potatoes [30].
Therefore, the goal of our work reported here was to characterize the rhizosphere microbial community of different Dendrobium species during the maturity stage. Our specific objective was to describe taxa associated with each Dendrobium species and determine which the environmental factors were related to microbial diversity and community composition. Specifically, we hypothesized that: (1) the dominant genus bacteria of Dendrobium and its comparison with other medicinal plants; (2) rhizosphere community composition will differ between three different Dendrobium; and (3) the hub flora can be obtained through WGCNA analysis.

Plant material and soil sampling
Dendrobium plants were artificially cultivated in the greenhouse of Anhui Tongjisheng Biotechnology Company, Lu'an, China. The original source was collected by the company from the wild after obtaining local permission. The protocols and the conditions of planting for the growth of protocorm-like bodies were described by our previous study [31]. From the grown plants of the three species of Dendrobium, two-year-old D. huoshanense, D. moniliforme and D. officinale were selected to provide eight replicates of each sample. The voucher specimens were authenticated by Professor Maoyun Yu from West Anhui University and deposited at Jiangsu Key Laboratory of Crop Genetics and Physiology in Yangzhou University, Yangzhou, China (Voucher number: 20A01, 20A02 and 20A03).
To obtain rhizosphere soil, plants were removed from flowerpots and large soil aggregates were removed by hand; soil firmly attached to roots was collected with sterile brushes and regarded as rhizosphere soil. The rhizosphere soil was sampled and sieved to remove plant debris. A part of the soil samples was put into sterile centrifuge tubes, frozen in liquid nitrogen immediately, and then stored at − 80°C until the soil microbial composition was analyzed.

DNA extraction, PCR amplification and 16S rRNA sequencing
For each flowerpot we obtained a single DNA sample, which was used for 16S rRNA sequencing. Total DNA extraction and concentration were used DNeasy Power Soil Kit (Qiagen, Valencia, CA, USA), Qubit® dsDNA Assay Kit (Life Technologies, CA, USA) and DNA degradation degree and potential contamination were monitored on 1% agarose gels [32].
For bacterial diversity analysis, V4 hypervariable regions of 16S rRNA genes were amplified with universal primers 515 F (5′-GTGCCAGCMGCCGCGG-3′) and 806 R (5′-GGACTACHVGGGTWTCTAAT-3′) [33], and then PCR products were sequenced by IonS5™XL sequencing technique platform [34]. Quality filtering on the raw reads were performed under specific filtering conditions to obtain the high-quality clean reads according to the Cutadapt (V1.9.1) [35] quality-controlled process. Chimeric sequences were removed by comparing with the Silva database [36] using UCHIME algorithm [37] to detect chimera sequences, and then the chimera sequences were removed. Operational taxonomic units (OTUs, cutoff 97% sequence identity for 16S rRNA) were clustered by UPARSE software (v7.0.1001) [38]. The phylogenetic taxonomy was assigned according to the Ribosomal Database Project (RDP) classifier at an 80% confidence threshold (Version 2.2) [39] using the Silva databases for bacteria. Alphadiversity was described for each sample using the following metrics for the observed species (OTU numbers); Chao1, Shannon and Simpson index, ACE and Good-coverage, which were generated to compare the level of bacterial OTU diversity [40].

WGCNA analysis Construction of weighted OTUs co-expression networks and identification of modules
The co-expression network of OTU was constructed using the WGCNA program package in R software [41]. Co-expression module construction was according to Wan et al. methods [42]. We used the function (pick-SoftThreshold) to calculate the appropriate weighting coefficient β. The selection of the soft threshold should satisfy that the square of the correlation coefficient is close to 0.8 while ensuring that the connectivity cannot be too low. The WGCNA algorithm was used to construct the modules to extract the OTU information corresponding to each module after the soft threshold is determined.

Interactions analysis of co-expression modules
According to the soft threshold selected above, the adjacency matrix was calculated, and the topological overlap matrix (TOM) was further constructed. The value in the TOM matrix reflects the similarity of the co-expression relationship between two OTUs. The higher the similarity, the closer the value is to 1. Then, we used the function (flashClust) to perform hierarchical clustering and used the dynamic cut tree algorithm (minModuleSize = 25) to cut the gene cluster tree for module division [41]. At the same time, colors were randomly assigned to each module, and principal component analysis was used to calculate the module feature vector of each module. The module feature vector was used to estimate the relationship between the module and the trait, the samples were classified according to the corresponding characteristics, and the module with the correlation coefficient ≥ |0.5| and p value ≤0.01 was selected for further analysis.

Identification of hub OTUs
Hub OTUs are often used as an abbreviation for highly connected OTU, which has a high degree of connectivity in the co-expression module. In our research, according to the size of the module, we classified the top 20 OTUs with the highest correlation as hub OTUs. At the same time, we used Cytoscape (v.3.6.1) to build and visualize the interaction network between OTU and OTU [43].

Statistical analysis
The WGCNA software package, stat and ggplot2 packages in R software (Version 2.15.3) were used for PCoA and WGCNA analysis. We conducted the co-occurrence network analysis in different rhizosphere soil compartments. The network analyses were performed using the psych R package (v.1.8.12) [44]. The bipartite network was constructed to feature the shared phyla among different samples [45]. The bipartite network was visualized using Cytoscape (v.3.6.1) [43].
Results αand β-diversity of soil microbial communities The rhizosphere soil alpha diversity indexes of different Dendrobium species are different (Fig. 1). The Good's coverage index of Dendrobium library was greater than 98.5%, indicating that the sequencing results reflect the real situation of the bacterial population in the sample. We found that Dendrobium huoshanense (Dh) had the highest Shannon diversity index (10.29), while Dendrobium officinale (Do) had the lowest (9.16), which indicated that Dh rhizosphere soil has the highest diversity of bacterial communities. The index of ACE and Chao1 have the same trend, Dh is the highest, respectively 5225 and 4877, indicating that Dh samples have the highest community richness. Further analysis revealed that the Shannon index and Simpson index of soil bacteria of Dh was significantly different from Do, and Do was significantly different from Dendrobium moniliforme (Dm).
Principal coordinate analysis (PCoA) was performed at the operational taxonomic unit (OTU) level (Fig. 2). The analysis of PCoA using the weighted UniFrac distance, indicated a distinct pattern in the rhizosphere bacterial communities associated with the two axes, explaining 29.58 and 19.62% of the total variation in the Dendrobium rhizosphere soil. Using the unweighted UniFrac distance, the two axes explained 17.64 and 11.79% of the total variation in the Dendrobium rhizosphere soil. Bacterial communities associated with the rhizosphere soil were clustered in three regions according to a PCoA using the unweighted UniFrac distance, corresponding to Dh, Do, and Dm. However, no clustering was detected using the weighted UniFrac distance from different Dendrobium species.

Taxonomic classification and abundance
Rarefaction curves for bacterial communities suggested that changes in OTU density within the different Dendrobium species was sufficiently captured, and the sequencing was relatively comprehensive in covering the microbial communities (Fig. 3a). After quality filtering and processing according to a 97% similarity, a total of 240,320 sequences were obtained from these three Dendrobium species, respectively: 80,101 for Dh, 80,109 for Do and 80,110 for Dm. The OTUs common to the different Dendrobium species are presented in Fig. 3b as a Venn diagram to assess the relationships among the bacterial communities. 5609 OTUs were shared by Dh, Do and Dm; 1041 OTUs were shared by Dm and Do; 926 OTUs were shared by Dh and Do; and 971 OTUs were shared by Dm and Dh. The numbers of OTUs unique to each species were as follows: 1007 for Dm, 909 for Do, and 716 for Dh. A total of 11,179 OTUs were detected in all samples (Fig. 3b).
According to the analysis of OTU annotation results, different Dendrobium rhizosphere soil bacteria include 2 kingdoms, 63 phyla, 72 classes, 159 orders, 309 families, 850 genera and 663 species (Fig. 4). Among all sequences, the dominant bacterial phyla (relative abundance > 1%) were Proteobacteria, Actinobacteria, Bacteroidetes, Acidobacteria, Firmicutes, Verrucomicrobia, Planctomycetes, Chloroflexi, and Gemmatimonadetes, with contributions of 45.41, 17.58, 9.94, 7.77, 3.93, 3.03, 2.88, 1.98 and 1.14%, respectively. The difference at the phylum level of different Dendrobium species is not large, but there were more and more obvious differences in the level of the class afterwards. At the order level, the relative abundance of Alteromonadeles on Dh and Dm was low, while the relative abundance on Do was relatively high. Pseudoalteromonadaceae at the family level and Pseudoalteromonas at the genus level also have the same relative abundance pattern. At the species level, although the top ten species of bacteria in the rhizosphere soils of the three Dendrobium species were the same, the relative expression abundances were very different.

LEfSe analysis and co-occurrence patterns between rhizosphere soil microbiomes
To further elucidate the possible interactions between identified bacterial dependencies in rhizosphere soil samples, linear discriminant analysis (LDA) and effect size (LEfSe) methods were used for quantitative analysis of biomarkers in different species. We detected significant differences in the abundance of bacterial biomarkers from different groups and identified a total of 13 biomarkers from all rhizosphere soil samples as shown in the branching diagram (Fig. 5b). The significant taxa in the Dh were affiliated with diverse phylogenetic groups, including the family Sphingomonadaceae, and order Sphingomoadales. In the Dm, the significantly abundant taxa were the genus Flavobacterium, the family Flavobacteriaceae and the order Flavobacteriales. In the Do, the significant taxa belonged to the phylum Actinobacteria, class unidentified Actinobacteria and Acdobacteriia, orders Pseudonocardiales and Acidobacteriales, family Pseudonocardiaceae and unidentified Acidobacteriales, and genus Crossiella, which were all abundant.
Network analysis is widely used in the exploration of complex microbial communities. In the rhizosphere soil of Dendrobium, the distributions and abundances of microbial genera are significantly different. In order to further explore the interaction between microorganisms in the root micro-ecological environment of Dendrobium, the contributions of different microbial populations to the overall community structure were evaluated using a bipartite association network to visualize the associations between the phylum and rhizospheric compartments (Fig. 7). In the bacterial community, 7 clusters of phyla were generated in the three kinds of Dendrobium species of the bipartite network. In the middle cluster of the bacterial community, 37.5% (3687) of the indicator phyla were associated with all the Dendrobium species, which is a large proportion of the indicator phyla.

Construction of gene co-expression network
We used the WGCNA program package in R software to construct the co-expression network and selected 11, 179 OTUs from the sequencing results of the rhizosphere soil microbiome of 24 Dendrobium samples to construct the co-expression network. To begin with, we loaded the R package, read the expression value matrix, calculated and clustered the correlation coefficients of the expression levels of each sample, and drew a heat map of species characteristics. As shown in Fig. 5a, most of the samples are clustered into one group according to species. The relative abundance of OTU is ranked as Dendrobium moniliforme (Dm) > Dendrobium officinale (Do) > Dendrobium huoshanense (Dh). In order to make the co-expression network meet the condition of nonscale distribution, we needed a suitable weighting coefficient β; therefore, the function (pickSoftThreshold) was used to calculate β. While satisfying that the square of the correlation coefficient is close to 0.8, it was ensured that the average connectivity could not be too low. According to these two conditions, we chose β = 26, and the square of the correlation coefficient was 0.85.
The soft threshold obtained above was used to construct a scale-free network. Firstly, the adjacency matrix was calculated based on the expression value matrix, and then the topological overlap matrix reflecting the common expression similarity was deduced. Then, the topological overlap matrix was used for hierarchical clustering to draw a cluster tree that could jointly express the overall distribution characteristics of similarity. Finally, the dynamic shear tree algorithm was used to cut the resulting cluster tree. In this process, OTUs with high similarity in common expression were gathered into the same branch, and different branches of the cluster tree represented different modules, and each module was given a specified color (Fig. 5b). Ten coexpression modules were obtained after constructing the co-expression network (Table 1). Among them, turquoise module had the largest number of OTUs, which is 501, while purple module had the smallest number of OTUs, which was 29.

Interaction analysis of co-expression module
In order to find the interaction relationship between coexpressed network modules, we analyzed the correlation of eigenvectors. Firstly, cluster analysis of eigenvectors was carried out (Fig. 6a). These 10 clusters were divided into two large clusters, including four modules (Brown, Black, turquoise and Red) and the other six modules. Therefore, the correlation between different modules was also significantly different, as shown in Fig. 6b. We analyzed the correlation between module feature vectors and species characteristics. As shown in Fig. 6c, four modules are highly correlated with species characteristics: turquoise module (p-value =7.5× 10 − 5 , correlation coefficient = − 0.731), Red module (p-value =0.0096, correlation coefficient = − 0.528), yellow module (p-value =0.00028, correlation coefficient = 0.688) and green module (p-value =0.00033, correlation coefficient = 0.682). Therefore, these four modules were further studied as modules of interest below.

Hub OTUs identification
From the above study, we obtained four modules with the most significant differences in species  Additional file Table 1), we have carried out network analysis on these four modules (Fig. 7)

Discussion
Rhizosphere microorganisms can co-exist with plant roots, colonize and be maintained in roots, and play an important role in promoting plant growth and development [46]. Among these microorganisms, the utilization rate and sensitivity of bacteria to root exudates are far higher than that of fungi, and bacteria are the most active and dominant microorganisms in rhizosphere [47,48]. In addition, the number and species of rhizosphere microorganisms have a direct impact on soil biochemical activity and nutrient transformation. Under the influence of various complex factors of natural conditions, there are great differences in rhizosphere microbial flora of different plants and even different genotypes of the same plant [49,50]. So, in this study, bacterial communities associated with the rhizosphere of three different Dendrobium species were characterized by highthroughput sequencing during the maturity stage.

Bacterial communities in rhizosphere soil of Dendrobium
It can be seen from the results that the dominant phyla in the rhizosphere soil of Dendrobium are Proteobacteria, Actinobacteria, Bacteroidetes, Acidobacteria, Firmicutes, Verrucomicrobia, Planctomycetes, Chloroflexi and Gemmatimonadetes. This indicates the richness and diversity of the rhizosphere microbial community of Dendrobium. However, the main dominant bacteria of Dendrobium are still concentrated in Proteobacteria, Actinobacteria and Bacteroidetes. Lundberg et al. found that the relative abundance of Proteobacteria, Bacteroidetes, Actinobacteria, Acidobacteria, Firmicutes and Gemmatimonadetes in Arabidopsis rhizosphere soil were relatively high, and were the dominant bacteria in Arabidopsis rhizosphere soil [8]. Ling et al. studied the rhizosphere microorganisms of watermelon, and found that Acidobacteria, Actinomycetes, Bacteroidetes, Cyanobacteria, Firmicutes and Proteobacteria were the main leading bacteria [51]. Davide et al. found that Actinobacteria, Bacteroidetes and Proteobacteria were dominant in barley rhizosphere soil [52]. Although there are differences in the bacterial communities of different plant rhizosphere soils, the dominant phyla of Actinobacteria, Bacteroidetes and Proteobacteria are the common dominant phyla of Dendrobium and the abovementioned plants, indicating that these phyla may be the common dominant phyla of all plant rhizosphere bacterial communities. Proteobacteria predominates in all ecosystems, especially in soil systems [53,54], due to the fact that Proteobacteria contains a large level of physiological, morphological and metabolic diversity, and that Proteobacteria is of great significance to the C and N cycles [55]. Proteobacteria reproduce fast, have good adaptability to unstable carbon sources, and are widely distributed in the global soil environment [56]. In general, the abundance of Proteobacteria or Acidobacteria in soil samples is the greatest. These bacterial  c Module-trait associations. Each row corresponds to a module characteristic gene (eigengene), and each column corresponds to a trait. Each cell contains a corresponding correlation and p-value. According to the color legend, the table is color-coded by correlation groups rich in rhizosphere microorganisms of Dendrobium, are also found to be the dominant communities of other plant rhizosphere microbiomes [57,58].
Differences of bacterial communities in rhizosphere soils of three different Dendrobium species According to the diversity indices alpha and beta, there were significant differences among the three Dendrobium species. The relative abundance of Acidobacteriales, Pseudonocardiaceae, Pseudoalteromonas and Pseudomonadales in Dendrobium officinale were higher than those in the other two species. Acidobacteriales is the dominant bacteria in the common plant rhizosphere bacterial community. Because Acidobacteriales can degrade complex root exudates such as cellulose and lignin, it plays a major role in the plant rhizosphere carbon cycle [59]. Pseudoalteromonas sp. secretes a variety of extracellular active substances, including proteins, polysaccharides, brominated compounds, extracellular enzymes, extracellular toxins, antibiotics and so on [60]. These substances have antibacterial, algicidal, bactericidal and cellulose degrading activities [61,62]. Pseudomonadales is an important group of biocontrol microorganisms, and is also one of the most widely distributed microorganisms in nature. Its rapid reproduction, strong colonization ability and simple nutrition requirements have been widely studied for its inhibition of plant diseases and promotion of plant growth.
In the existing studies, Pseudonocardiaceae is mainly associated with cellulose degradation and antibiotic synthesis [63][64][65]. It can be seen from the above that the function of the bacterial community in the rhizosphere of Dendrobium officinale may be stronger than that of the other two kinds of Dendrobium. In Dendrobium moniliforme, the relative abundance of Bacteroidetes is relatively high, and Bacteroidetes is a poor nutrient bacterium, which is suitable for growth in the environment with less absorbable nutrients such as organic matter and available nitrogen [66].
In bacterial species, we found that Lysobacter soli have a high relative abundance in Dendrobium moniliforme. This fungus has previously been isolated from the soil where ginseng is grown [67], and has been found in other plants to promote plant activity [68]. Psychrobacter ibarius has been found in Dendrobium officinale. Some researchers have isolated this bacterium from the root plane of Angelica sinensis [69]. This is a fungus related to polysaccharide synthesis. This is consistent with our previous research results [13]. Our previous research also found that among these three kinds of Dendrobium, the polysaccharides of two-year-old Dendrobium are higher than those in the other two kinds of two-year-old Dendrobium.
WGCNA applied to the rhizosphere microbiome in Dendrobium species In recent years, with the development of highthroughput sequencing technology, network analysis based on high-throughput omics data has become a popular big data processing method due to its high efficiency and convenience [70]. WGCNA, as one of the methods of network analysis, can specifically screen OTUs highly correlated with the target traits, divide them into multiple co-expression modules, and mine the core OTUs, which has been proved to be a fast, accurate and efficient method for data mining in potatoes and other plants [71,72]. In order to understand the differences in the rhizosphere microbial communities of Dendrobium, this study used the WGCNA package in the R software to construct a co-expression network of the rhizosphere microbial community of Dendrobium, and explored the community structure of the rhizosphere microbes in a new way. A total of 10 modules were identified in the constructed Dendrobium rhizosphere microbial co-expression network, and the 4 modules with the highest correlation with species characteristics were further analyzed. Each module obtained 20 hub OTUs. By annotating these hub OTUs, the dominant bacteria phyla in the rhizosphere microbial community of Proteobacteria, Acidobacteria and Actinobacteria were finally obtained.
Compared with the results of this study, a total of 22 co-expression modules were identified in the rhizospheric co-expression network in the WGCNA analysis of rice root-associated microbiomes. By analyzing the OTUs annotations of the three microbial co-expression networks with the most significant differences among the three niches within the endosphere, rhizoplane and rhizosphere, it is concluded that the hub bacteria phyla in the rice rhizosphere microbial community is Proteobacteria, Actinobacteria and Bacteroidetes, which are highly similar to the dominant bacteria phyla we found in the rhizosphere microbial community of Dendrobium.
The dominant microorganisms are relatively abundant in the soil and play an important role in the regulation of ecological functions [73]. Proteobacteria and Actinomycota are enriched in rhizosphere and non-rhizosphere environments, which is related to their strong adaptability. Among them, the Proteobacteria can survive and multiply in most environments and become the dominant bacteria because its outer membrane is mainly composed of lipopolysaccharide, which can protect its internal genetic material from external interference [74,75]. In addition, the phylum Proteobacteria plays an important role in adapting to environmental changes and resisting adversity stress [76]. Actinomycota plays an important role in the decolorization of the toxic dyetriphenylmethane (TPM) in the soil [77], and plays a certain role in adaptation to arid environments [78]. Therefore, we speculate that the Proteobacteria and Actinomycota play a pivotal role in the plant rhizosphere microbial community, and they play an important regulatory role in the plant rhizosphere.

Conclusions
Overall, the chemical properties of rhizosphere soil of Dendrobium and the climatic characteristics of Dendrobium cultivation were firstly analyzed. Then, combined with high-throughput sequencing technology, the rhizosphere soil microbial communities of different species of Dendrobium were studied, and the rhizosphere microorganisms and environmental driving factors of different species of Dendrobium were discussed. The main conclusions are as follows: the dominant bacteria in rhizosphere soil of Dendrobium are as follows: Proteobacteria, Bacteroidetes, Actinobacteria, and Acidobacteria. These are the dominant bacteria in the rhizosphere bacterial community. However, there are some differences in the bacterial communities among different Dendrobium species, and the smaller the bacterial level is, the greater the difference is. Through WGCNA analysis, it is found that in the rhizosphere microbial community network of Dendrobium, the hub flora is also in the Bacteroides, Acidobacteria and Proteobacteria.

Funding
This project was funded by the Priority Academic Program Development of Jiangsu Higher Education Institutions (PAPD), the National traditional Chinese medicine standardization project (ZYYS-2014 [12]). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Availability of data and materials Raw amplicon sequence data related to this study were deposited in the NCBI Sequence Read Archive (NCBI SRA) under Bioprojects PRJNA638443.
Ethics approval and consent to participate Dendrobium huoshanense, Dendrobium officinale and Dendrobium moniliforme used in this study were cultivated by Maoyun Yu's greenhouse from Anhui Tongjisheng Biotechnology Company, Lu'an, China. No permits were required for the collection of the samples.

Consent for publication
Not applicable.