Network Analysis on Bacterial Co-Occurrence Patterns in a Eutrophic Lake across a Perturbation Event

Network Analysis on Bacterial Co-Occurrence Patterns in a Eutrophic Lake across a Perturbation Event

Corresponding author: Nan Li, Florida A& M University, 1515 S Martin Luther King, Jr Blvd, Tallahassee, FL 32307, USA. Email: nli0417@gmail.com, nan.li@famu.edu

Abstract

Several studies have assessed the variation and stability of bacterial communities in lakes in response to perturbation, yet spe- cificecological relationships among taxa affected by the perturbation are largely unknown. In this study, we used molecular eco- logical network (MEN) analysis to characterize co-occurring OTUs patterns across a major perturbation event (i.e. drainage) in Lake Munson. The OTUs within phyla tended to co-occur in a range from 10% to 76%. OTUs belonging to Acidobacteria, Actino- bacteria, Bacteroidetes, Chloroflexi, Cyanobacteria and Nitrospira showed strong co-occurrence. Co-occurrence patterns were observed to remain relatively stable across the perturbation event. The results revealed putative inter-taxa correlations within bacterial communities across a perturbation event and advanced the understanding of their structureand specifice cologicalroles.

Keywords: Networks Analysis; Co-Occurrence; 16S Rrna; Freshwater; Pyrosequencing

Introduction

Network analysis has been widely applied to explain the functional organization of various biological systems and to reveal the interactions of biological molecules, including genes and proteins [1-3]. Recently, this network concept was used to investigate inter-taxa correlations of microbial taxa in a 16S rRNA gene pyrosequencing dataset of soil and water samples to demonstrate general non-random associations at broad taxonomic levels [4-6].

Network analysis of taxa co-occurrence patterns offers new insights into the structure of complex microbial communi- ties, and expands the information provided by more stan- dard analytical approaches such as correlation analysis.

Several studies have assessed the variation and stability of bacterial communities in lakes in response to perturbation [7-10]. For example, a study on Lake North Sparkling Bog showed a change in microbial community composition after artificial mixing of the lake, and that some bacterial commu- nities were able to return to their pre-manipulation state when certain environmental conditions, such as dissolved oxygen, were restored [10]. Recently, the association net- work approaches have been used to assess bacterial inter- dependencies in a temporal survey of freshwater bacterial groups from Lake Erken by an annual sampling [6]. However, differences in bacterial interactions within a lake across per- turbation an event have not been well examined.

In this study, we used molecular ecological network (MENs) analysis to characterize prokaryotic taxa associations in Lake Munson, a eutrophic freshwater lake located in Tal- lahassee, FL, USA across a major perturbation event. In an effort to improve its water quality, Lake Munson was drained from October 2010 to May 2011, offering a unique opportu- nity to explore co-occurrence patterns of freshwater bacterial communities across drainage, which represented a major per- turbation. Water samples from the lake were collected season- ally. Over 40,000 bacterial 16S rRNA gene sequences were an- alyzed to investigate the following question: Which taxa and/ or species compose the core part of the co-occurrence patterns (broadly distributed and strongly correlated) in the bacterial community of Lake Munson across its drainage?

Materials and methods

Sampling

Before drainage of Lake Munson (N 30°21.923’; W 84°18.091’), water samples were collected on March 4, 2010, June 8, 2010,

August 8, 2010, and October 15, 2010 (Fig. 1). A Niskin water sampler was used to collect 500 ml of water below the surface from each side of the boat and immediately combined to form a composite sample of 1 L. There are no specific permissions were required for these locations/activities from Lake Mun- son.

Figure 1. Lake Munson Sampling scheme.

As the lake began to refill following drainage, water samples were collected from the boat ramp on August 25, 2011, Sep- tember 16, 2011, and November 18, 2011 as the water level in the lake was too low to take a boat out for sampling. Following collection, all samples were stored on ice and transported to the laboratory for processing. There the samples were filtered through 0.2-μm Nuclepore track-etched membrane filters (What man) to capture the bacterial community and the fil- ters were stored at -20°C for further molecular analysis.

During drainage of the lake, sediment samples were collected in triplicate with a sediment corer on December 9, 2010, two months after the lake was drawn down and the sediment was still wet. On March 9, 2011, five months after drainage of the lake had begun and the lake was relatively barren of water, sed- iment samples were collected in triplicate from the same site location as the December sampling. Sediment samples were also collected on May 20, 2011, a few weeks before the dam was scheduled to close and the lake began to refill. However, by this time, the original sample site was covered in waist-high grass making it inaccessible. Therefore, triplicate sediment samples were collected from a site near the center of the lake

that had no grass and very little water, allowing easier collec- tion with a sediment corer. All samples were placed on ice for transport to the laboratory where they were stored at -20°C until further processing.

DNA extraction and pyrosequencing

The genomic DNA from the before and after drainage samples were extracted from the bacteria captured on the filters using the Mo-Bio Power Water® DNA Isolation Kit (Carlsbad, CA) following the manufacturer’s instructions. Sediment genom- ic DNA was extracted using Power Soil DNA isolation kit (Mo Bio Laboratories, Carlsbad, CA) following the manufacturer’s instructions. DNA yield and purity were measured via a micro volume fluorospectrometer (Nano Drop Technologies, Dela- ware).

Preparation for pyrosequencing entailed amplifying the 16S rRNA genes in the sediment and water samples with the uni- versal primers 27F and 1492R 11]. Two μL of template DNA was aliquoted into illustra™ PuReTaq Ready-To-Go™ PCR Beads (GE Healthcare, Waukesha, WI), with PCR grade molecular wa- ter, and 5 pmol/μL of each primer for a total reaction mixture of 25 μL. PCR was conducted on a Biorad thermocycler (Hercu- les, CA) and conditions were as follows: 3 min initial denatur- ation at 95°C, 35 cycles of 30 s denaturation at 95°C, 1 min an- nealing at 55°C, 2 min elongation at 72°C, and a final extension for 7 min at 72°C. Product size and purity were confirmed by electrophoresis in 0.7% agarose gels using ethidium bromide. Amplicons were used for 454 pyrosequencing with a 454 GS FLX system (454 Life Sciences) according to manufacturer’s instructions at the Research and Testing Laboratory LLC. In Lubbock, TX (USA). 454 pyrosequencing reads were deposited into the NCBI Short Read Archive under accession SRP022185.

Taxonomic analysis

Pyrosequencing data was analyzed using the single software platform, mothur v.1.26.0 [12]. This platform allows sequenc- es to be trimmed, screened, aligned, assigned to operational taxonomic units (OTUs) and classified. For the phylotype-inde- pendent approach, sequences were clustered into operational taxonomic units (OTUs) at a distance threshold of 0.03 (97% similarity) via the average neighbor algorithm [13].

To minimize the effects of random sequencing error, low qual- ity sequences were eliminated as described by Schloss et al. [14]. After removal of barcodes and primers, the remaining se- quences were trimmed so that all of them started and ended at similar positions in their alignment to the SILVA database and underwent screening for chimeras through UCHIME [15]. After removal of chimeras, sequences were classified using the Ribosomal Database Project (RDP) Naïve Bayesian Classifier (minimum confidence of 50%), which provides taxonomic as- signments from domain to genus based on 16S rRNA sequenc- es [16]. After assigning taxonomy, contaminants such as chlo-

roplasts and mitochondria were removed to further improve data quality.

Network analysis

MENs were constructed as described by Ye Deng et al. (2012). The OTUs that occurred only once in the different samples were not included in the analysis. The number of sequences in the individual OTUs varied significantly between the different samples, therefore, the relative proportions of sequence num- bers were used for Pearson correlation analysis. Subsequently, a similarity matrix was obtained by taking the absolute values of the correlation matrix. An appropriate threshold for de- fining network structure, similarity thresholds (st), was de- fined using the random matrix theory (RMT)-based network approach [17], which encodes the strength of the connection between each pair of nodes. The sub-modules within a large module were detected by fast greedy modularity optimization [18]. For network comparison, random networks correspond- ing to all phylogenetic molecular ecological networks (pMENs) were generated using the Maslov-Sneppen procedure [19]. The networks were explored and visualized by Software Cyto- scape 2.8.3 [20].

Discussion

Network description

Taxonomic analysis revealed that Proteobacteria was the dominant phylum, representing at least 35% of total bacteri- al sequences in each sample (Figure S1). Among the Proteo- bacteria, Beta-proteobacteria was the dominant class in all of the libraries. The second most abundant phylum was Bacteri- odetes, representing 10%–25% of bacterial sequences in each sample (except for one time point, D-03.09.2011). Consistent with previous studies [7,21,22] ,differences were observed in the microbial community structure between water and sed- iment samples. For example, acidobacteria are mainly found in the semimetal samples, but fewer detected in the water sam- ples; and cyanobacteria are more dominate in the water sam- ples then its in the semimetal samples(Fig. S1).

It is well-known that in natural environments, individual or- ganisms typically do not function independently, but rath- er forms a complex inter-species interaction system [23] for activities such as predation [24] and competition [25]. These interactions shape the structure of the ecological communi- ty and play an important role in species evolution [26-28]. In comparing samples obtained before and after drainage of Lake Munson, it is obvious that the majority of microbial community species had recovered (Figure S1). For example, the percentage of actinobacter varied from 28% at pre drain- age (Figure S1, Pre-D-03.042010), reduced to around 2% (D-03-09-2011) , finally returned to 24% (PD-12.09.2011). Rate of cyanobacteria experienced a similar changing pat- tern, they were occupied 5% in the first water sample

(Pre-D-03.04.2010), reduced to less than 1% (D-03-09- 2011), and finally recovered to around 5% (PD-12.18.2011). This suggests that the microbial communities in Lake Munson are resilient even after exposure to a large perturbation event. The co-occurrence patterns for correlation among taxa and/ or species across the entire perturbation (before, during and after the drainage) remains illusive.

Figure S1 Distribution of phyla for each data library. Abbreviations: Pre-D, pre drainage; D, drainage; PD, post Drainage.

After preprocessing of the raw sequences, the numbers of se- quences in the 10 samples ranged from 1,775 to 6,853. After defining OTUs with a 0.03 cutoff, 8,740 distinct OTUs across the 10 samples were obtained. For each data set, only the OTUs that appeared in seven or more samples were used for correla- tion calculations, resulting in 244 OTUs in total for the data set. After threshold scanning by the RMT based approach, the pMENs were constructed with an identical similarity thresh- old of 0.57(P < 0.05). R2 values of the linear relationship be- tween logarithms of clustering coefficients and the logarithms of connectivity ranged from 0.10 to 0.83, indicating the hierar- chical behavior was quite variable. In MENs, a module in the network is a group of OTUs that are highly connected among themselves, but have much fewer connections with OTUs out- side the group [29]. Four (modularity values 0.300) and two modules (modularity values 0.292) were identified by the sim- ulated annealing approach [30] and by the greedy modularity optimization method [31], respectively. These results indicate that the co-occurrence of OTUs across the perturbation were strongly connected to each other, but less so between groups. Finally, our constructed MEN before, during and after the per- turbation of Lake Munson exhibited a scale-free, small world with a variety of modularity properties, although hierarchical properties were only displayed in certain networks (Figure 2).

Figure 2 Network of co-occurring OTUs based on correlation analysis (P-value < 0.01). A triangle indicates a strong correlation (Coefficient

> 0.6), while a circle indicates a moderate correlation (coefficient < 0.6).

The topological roles of the OTUs

The topological roles of nodes in the drainage pMENs were il- lustrated by a Z-P plot (Fig. 3). Since different nodes have dis- tinct topological roles in the network, the analysis of modular topological roles can show the key populations based on the nodes’ roles [4,32]. Their topological roles can be defined by two parameters, within-module connectivity (Zi) and among module connectivity (Pi). According to the values of Zi and Pi, the roles of nodes were classified into four categories: periph- erals, connectors, module hubs and network hubs[4]. From an ecological perspective, peripherals might represent specialists, whereas connectors and module hubs may suggest generalists and network hubs may represent super generalists.

Figure 3 ZP-plot showing distribution of OTUs based on their mod- ule-based topological roles. Each dot represents an OTU in the Data- set from Lake Munson. The topological role of each OTU was deter- mined according to the scatter plot within-module Connectivity (z) and between-module connectivity (P).

The majority of OTUs (79.5%) identified across the three sam- pling periods were peripherals, with most of their links within their own modules. A total of 16 nodes (17.2 %) were connec- tors and only four nodes (3.3%) were module hubs. The three

OTUs that classified as module hubs were from Planctomycetes, Rhizobiales and Bacteroidales, indicating that they may be the representative species for the lake. No network hubs (super generalists) were observed. In the connector groups, 2 OTUs were assigned to Aquificae and Planctomycetes, while the oth- ers belonged to various taxonomic groups (i.e Actinobacteria, Bacteroidetes, Proteobacteria and Verrucomicrobia), generally common species in lake bacterial communities [10,21].

Interestingly, some bacteria with a specific role in the decom- position of organic materials were classified into the connector group (generalists) in the bacterial community network. They may well have become adapted to the eutrophic condition in Lake Munson. Two such examples are Pleomorphomonas ko- reensis and Methylobacter bovis which have been reported to produce key enzymes involved in the primary pathways of the nitrogen-fixing cycle [33-35]. Collectively, these results sug- gest the co-occurrence of microbial taxa were existed in Lake Munson across the drainage event. Further the microbial taxa can be classified into distinct topological roles which may indi- cate they perform similar ecological roles.

Co- occurrence patterns

The networks revealed not only the occurrence of patterns among tribes, but also associations among and between mem- bers of the bacteria community [6]. The resulting microbial network consisted of 93 nodes and 933 edges with an aver- age node connectivity of 24.55. The average network distance between all pairs of nodes (average path length) was 2.067 edges with a diameter (longest distance) of 47 edges. The av- erage clustering coefficient – that is, how nodes are embedded in their neighborhood and the degree to which they tended to cluster together – was 0.504. These structural properties can potentially be used for quick and easy comparisons be- tween complex datasets from different ecosystem types to explore how the general traits of a certain habitat type may influence the assembly of microbial communities. The nodes in these networks belonged to 13 phyla. The distribution of these nodes was different from the basic taxonomic analysis, with Acidobacteria, Actinobacteria, Bacteroidetes, Chloroflexi, Cyanobacteria and Nitrospira being the most dominant. The co-occurrence of OTUs from the same phyla ranged from 10% to 76% (Figure 4).

The structural analysis of the networks (Figure 2) in combina- tion with statistical methods such as correlation analysis can also be used as an index of the synergetic relationships in an ecological niche. For example, the node belonging to Clostridia- ceae only showed interactions with Caldilineaceae, Bacillaceae and Desulfurobacteriaceae. To show the strong connection be- tween the community members, a sub-network was defined if the clustering coefficient was larger than 0.60 [4,29]. The OTUs belonging to Chlorobi, Cyanobacteria, Planctomycetes and Ver- rucomicrobia did not show strong co-occurrence across the

Table 1 Taxonomy of OTUs (triangle nodes in Figure 1) and their positive co-occurrent OTUs

Cosmopolitan Co-occurring Phyla Coefficient Number of co- Closest relative
Phyla occurring OTUs species
Acidobacteria Actinobacteria 0.63 5 Conexibacter
Aquificae 0.65 1 Thermovibrio
Chloroflexi 0.70 2 Leptolinea
Proteobacteria 0.73 18 Ancylobacter
Verrucomicrobia 0.50 1 Roseibacillus
Actinobacteria Aquificae 0.65 1 Thermovibrio
Bacteroidetes 0.74 7 Arcicella
Chloroflexi 0.70 1 Leptolinea
Firmicutes 0.61 1 Caldalkalibacillus
Nitrospira 0.66 1 Nitrospira
Proteobacteria 0.64 24 Polynucleobacter
Verrucomicrobia 0.61 2 Cerasicoccus
Aquificae Chloroflexi 0.70 3 Leptolinea
Firmicutes 0.62 1 Caldalkalibacillus
Nitrospira 0.66 1 Nitrospira
Proteobacteria 0.65 12 Pelagibacter
Bacteroidetes Proteobacteria 0.64 3 Polynucleobacter
Verrucomicrobia 0.61 1 Cerasicoccus
Chloroflexi Firmicutes 0.83 2 Clostridium
Nitrospira 0.66 1 Nitrospira
Planctomycetes 0.52 1 Pirellula
Proteobacteria 0.66 14 Thioflavicoccus
Verrucomicrobia 0.61 2 Cerasicoccus
Firmicutes Nitrospira 0.66 1 Nitrospira
Planctomycetes 0.52 1 Pirellula
Proteobacteria 0.64 6 Cardiobacterium
Gemmatimonadetes Proteobacteria 0.66 6 Thioflavicoccus
Verrucomicrobia 0.50 1 Roseibacillus
Nitrospira Proteobacteria 0.66 9 Thioflavicoccus
Planctomycetes 0.52 1 Pirellula
Verrucomicrobia 0.78 1 Subdivision3
Proteobacteria Alphaproteobacteria 0.73 7 Pseudolabrys
Betaproteobacteria 0.73 4 Methylophilus
Deltaproteobacteria 0.73 4 Geobacter
Gammaproteobacteria 0.66 5 Thioflavicoccus
Verrucomicrobia 0.61 2 Cerasicoccus

Abbreviation: OTUs, operational taxonomic units.

OTUs belonging to ChlorobiCyanobacteriaPlanctomycetes and Verrucomicrobia had no positive co-occurrent matches.

perturbation (Fig. 2 B). In contrast, the majority of strongly cor- related nodes belonged to Proteobacteria and Actinobacteria, which also showed substantial interaction with other microbi- al groups (edge larger than 5). Interestingly, some nodes that showed strong correlations did not connect with other nodes (Fig. 2 B), suggesting that they may be independent in the en- vironment, but restricted by basic environmental factors. An example is Holophaga, which thrives in anaerobic conditions and is thus more abundant in soil samples [36]. Additionally, co-occurrence patterns of the bacterial community across the various stages of perturbation in Lake Munson revealed that some species may play an important role in environmental functions. For example, Burkholderiaceae, detected in all sam- ples from before, during and after drainage of the lake, have been reported to be capable of removing heavy metals such as Cu, Pb, Cr, Ni, and Zn from contaminated water [37].

Figure 4. Relative abundances of different microbial taxonomic groups. A: number of co-occurrent OTUs. B: number of significant co-occurrent OTUs (nodes from Table 1).

Some of these co-occurrences may represent guilds of organ- isms performing similar or complementary functions to each other, while others may co-occur because of shared, preferred environmental conditions. When attempting to describe these collections, it is tempting to look for positive correlations among the phylogenetically related nodes in the sub-network that may indicate OTUs that are performing similar or com- plementary ecological roles as the correlation between Clos- tridiaceae and Caldilineaceae (Figure 2B). They were reported to have a capability to enhance biodegradation of petroleum liquids in the environment [38]. Petroleum liquids, referred to as Light Non-Aqueous Phase Liquid (LNAPL), pose a threat to the environment and human health [37]. Negative correlations may indicate competition or predation among the taxa, such as Chromatiaceae and Micrococcineae (Figure 2B),which have been reported to compete for sulfur in the process of sulfide oxidation in the environment [39-41]. And also a group of un- known species that were highly correlated with Cytophagace, suggesting possible symbioses or parasitism.

The analysis revealed a progression of the native microbial communities through a perturbation event. Overall, co-occur- rence patterns could not be simply explained by symbioses and/or other classical interaction relationships (eg. predation and competition). This may be due to the complexity of rela- tionships between phylogenetic and ecological components which may depend on the individual microbial group, as well as environmental conditions.

Conclusion

Previous studies on the biodiversity of lake microbial commu- nities during perturbation events have focused on the num- ber and abundance of species in the surface water and/or sediments [7-10]. To our knowledge, the interactions among species during the whole perturbation process have not been reported. Although we cannot claim that we have a compre- hensive view of interactions within the lake microbial com- munity, the results from this study showed that OTUs within phyla tended to co-occur in a range from 10% to 76%, and that OTUs belonging to Acidobacteria, Actinobacteria, Bacteroide- tes, Chloroflexi, Cyanobacteria and Nitrospira showed strong co-occurrence over the course of this study. Co-occurrence patterns were observed to remain relatively stable across the perturbation event. The current study provides a systematic approach to analyzing the structure and dynamics of bacterial communities during a major perturbation event in a freshwa- ter lake ecosystem.

Acknowledgement

This work was financially supported by the National Science Foundation HBCU-RISE Grant-0531523. There is no conflict of interest in this study.

Supplementary information is available at Jacobs Journal of Hydrology’s website.

References
  1. Duran-Pinedo AE, Paster B, Teles R, Frias-Lopez J. Correla- tion Network Analysis Applied to Complex Biofilm Communi- ties. Plos One. 2011, 6(12): e28438.
  2. Barabasi AL, Oltvai ZN. Network biology: Understanding the cell’s functional organization. Nature Reviews Genetics. 2004, 5(2): 101-U115.
  3. Chen S, Kim DK, Chase MW, Kim JH. Networks in a large-scale phylogenetic analysis: reconstructing evolutionary history of Asparagales (Lilianae) based on four plastid genes. PLoS One. 2013, 8(3): e59472.
  4. Zhou JZ, Deng Y, Luo F, He ZL, Yang YF. Phylogenetic Molec- ular Ecological Network of Soil Microbial Communities in Re- sponse to Elevated CO2. MBio. 2011, 2(4): e00122-11.
  5. Steele JA, Countway PD, Xia L, Vigil PD, Beman JM et al. Ma- rine bacterial, archaeal and protistan association networks re- veal ecological linkages. ISME J. 2011, 5: 1414-1425.
  6. Eiler A, Heinrich F, Bertilsson S.Coherent dynamics and asso- ciation networks among lake bacterioplankton taxa. The ISME journal. 2011, 6(2): 330-342.
  7. Muylaert K, Van Der Gucht K, Vloemans N, Meester LD, Gillis M et al. Relationship between bacterial community composi- tion and bottom-up versus top-down variables in four eu- trophic shallow lakes. Appl Environ Microbiol. 2002, 68(10): 4740-4750.
  8. Yannarell AC, Triplett EW. Within- and between-lake vari- ability in the composition of bacterioplankton communities: investigations using multiple spatial scales. Appl Environ Mi- crobiol. 2004, 70(1): 214-223.
  9. Haller L, Tonolla M, Zopfi J, Peduzzi R, Wildi W et al. Com- position of bacterial and archaeal communities in freshwater sediments with different contamination levels (Lake Geneva, Switzerland). Water Res. 2011, 45(3): 1213-1228.
  10. Shade A, Read JS, Youngblut ND, Fierer N, Knight R et al. Lake microbial communities are resilient after a whole-ecosys- tem disturbance. ISME J. 2012, 6(12): 2153-2167.
  11. Amann RI, Ludwig W, Schleifer KH. Phylogenetic identifica- tion and in situ detection of individual microbial cells without cultivation. Microbiol Rev. 1995, 59(1):143-169.
  12. Schloss PD, Westcott SL, Ryabin T, Hall JR, Hartmann M et al. Introducing mothur: open-source, platform-independent, community-supported software for describing and comparing microbial communities. Appl Environ Microbiol. 2009, 75(23): 7537-7541.
  13. Schloss PD, Westcott SL. Assessing and Improving Methods Used in Operational Taxonomic Unit-Based Approaches for 16S rRNA Gene Sequence Analysis. Applied and Environmen- tal Microbiology. 2011, 77(10): 3219-3226.
  14. Schloss PD, Gevers D, Westcott SL. Reducing the effects of PCR amplification and sequencing artifacts on 16S rRNA-based studies. PLoS One. 2011, 6(12): e27310.
  15. Edgar RC, Haas BJ, Clemente JC, Quince C, Knight R. UCHIME improves sensitivity and speed of chimera detection. Bioinfor- matics. 2011, 27(16): 2194-2200.
  16. Lan YM, Wang Q, Cole JR, Rosen GL. Using the RDP Classifier
  17. to Predict Taxonomic Novelty and Reduce the Search Space for Finding Novel Organisms. PLoS one. 2012, 7(3): e32491.
  18. Luo F, Yang Y, Zhong J, Gao H, Khan L et al. Constructing gene co-expression networks and predicting functions of un- known genes by random matrix theory. BMC Bioinformatics. 2007, 8: 299.
  19. Schuetz P, Caflisch A. Efficient modularity optimization by multistep greedy algorithm and vertex mover refinement. Phys Rev E Stat Nonlin Soft Matter Phys. 2008, 77: 046112.
  20. Maslov S, Sneppen K. Specificity and stability in topology of protein networks. Science. 2002, 296: 910-913.
  21. Cline MS, Smoot M, Cerami E, Kuchinsky A, Landys N et al. Integration of biological networks and gene expression data using Cytoscape. Nature Protocols. 2007, 2(10): 2366-2382.
  22. Jiang H, Dong H, Zhang G, Yu B, Chapman LR et al. Microbial diversity in water and sediment of Lake Chaka, an athalassoha- line lake in northwestern China. Appl Environ Microbiol. 2006, 72: 3832-3845.
  23. Briee C, Moreira D, Lopez-Garcia P. Archaeal and bacterial community composition of sediment and plankton from a su- boxic freshwater pond. Res Microbiol. 2007, 158(3): 213-227.
  24. Little AE, Robinson CJ, Peterson SB, Raffa KF, HandelsmanJ. Rules of engagement: interspecies interactions that regulate microbial communities. Annu Rev Microbiol. 2008, 62: 375- 401.
  25. Comte J, Jacquet S, Viboud S, Fontvieille D, Millery A et al. Microbial community structure and dynamics in the largest natural French lake (Lake Bourget). Microb Ecol. 2006, 52(1): 72-89.
  26. Signoretto C, Burlacchini G, Pruzzo C, Canepari P. Per- sistence of Enterococcus faecalis in aquatic environments via surface interactions with copepods. Appl Environ Microbiol. 2005, 71(5): 2756-2761.
  27. Losos JB, Leal M, Glor RE, De Queiroz K, Hertz PE et al. Niche lability in the evolution of a Caribbean lizard community. Nature. 2003, 424(6948): 542-545.
  28. Chen H, Athar R, Zheng G, Williams HN. Prey bacteria shape the community structure of their predators. ISME J. 2011, 5(8): 1314-1322.
  29. Zhou J, Xia B, Treves DS, Wu LY, Marsh TL et al. Spatial and resource factors influencing high microbial diversity in soil. Appl Environ Microbiol. 2002, 68(1): 326-334.
  30. Deng Y, Jiang YH, Yang Y, He Z, Luo F et al. Molecular eco-logical network analyses. BMC Bioinformatics. 2012, 13: 113.
  31. Guimera R, Amaral LA. Cartography of complex networks: modules and universal roles. J Stat Mech. 2005, P02001-1– P02001-13.
  32. Newman ME. Fast algorithm for detecting community structure in networks. Phys Rev E Stat Nonlin Soft Matter Phys. 2004, 69: 066133.
  33. Olesen JM, Bascompte J, Dupont YL, Jordano P.The modu- larity of pollination networks. Proc Natl Acad Sci U S A. 2007, 104: 19891-19896.
  34. Shishkina VN, Iurchenko VV, Romanovskaia VA, Malashen- ko Iu R, Trotsenko Iu A. [Alternativity of methane assimilation pathways in obligate methylotrophs]. Mikrobiologiia. 1976, 45: 417-419.
  35. Starostina NG, Pashkova NI, Tsiomenko AB. Detection and partial characterization of bacteriocin in the methanotrophic bacterium Methylobacter bovis. Biochemistry (Mosc). 1998, 63(10): 1122-1125.
  36. Im WT, Kim SH, Kim MK, Ten LN, Lee ST. Pleomorphomonas koreensis sp. nov., a nitrogen-fixing species in the order Rhi- zobiales. Int J Syst Evol Microbiol. 2006, 56(pt 7): 1663-1666.
  37. Anderson I, Held B, Lapidus A, Nolan M, Lucas S et al. Ge- nome sequence of the homoacetogenic bacterium Holophaga foetida type strain (TMBS4(T)). Stand Genomic Sci. 2012, 6(2): 174-184.
  38. Lobova TI, Feil EJ, Popova LY. Multiple antibiotic resistance of heterotrophic bacteria isolated from Siberian lakes subject- ed to differing degrees of anthropogenic impact. Microb Drug Resist. 2011, 17(4): 583-591.
  39. Zeman NR.Thermally enhanced bioremediation of LNAPL: Colorado State University.
  40. Shahnavaz B.Communauté bactérienne de sols alpins et fil- tres environnementaux: Université Joseph-Fourier-Grenoble I. 2009.
  41. Peduzzi S, Welsh A, Demarta A, Decristophoris P, Peduzzi R et al. Thiocystis chemoclinalis sp. nov. and Thiocystis cadag- nonensis sp. nov., motile purple sulfur bacteria isolated from the chemocline of a meromictic lake. International journal of systematic and evolutionary microbiology. 2011, 61(pt 7): 1682-1687.
  42. Friedrich CG. Physiology and genetics of sulfur-oxidizing bacteria. Advances in microbial Physiology. 1997, 39: 235-289.

Be the first to comment on "Network Analysis on Bacterial Co-Occurrence Patterns in a Eutrophic Lake across a Perturbation Event"

Leave a comment

Your email address will not be published.


*


Select Language