Search tips
Search criteria 


Logo of scirepAboutEditorial BoardFor AuthorsScientific Reports
Sci Rep. 2017; 7: 43483.
Published online 2017 March 6. doi:  10.1038/srep43483
PMCID: PMC5337950

Effect of predatory bacteria on the gut bacterial microbiota in rats


Bdellovibrio bacteriovorus and Micavibrio aeruginosavorus are Gram-negative proteobacteria that are obligate predators of other Gram-negative bacteria and are considered potential alternatives to antibiotics. Most studies focusing on predatory bacteria have been performed in vitro, thus the effect of predatory bacteria on a live host, including the impact on the ecology of the native microbiota, has yet to be fully examined. In this study, intrarectal inoculations of Sprague-Dawley rats with predatory bacteria were performed. Additionally, feces were collected for seven days post-inoculation to determine the effect on gut bacterial diversity. Rat colonic tissue exhibited no abnormal histopathological effects due to predatory bacteria. A modest increase in pro-inflammatory cytokines was measured in the colons of rats inoculated with predatory bacteria by 24 and 48 hours, with all but IL-13 returning to baseline by seven days. V4 16S rRNA gene sequencing of fecal DNA demonstrated minimal shifts in taxonomic representation over the week due to predatory bacteria. Changes in bacterial populations due to exposure to B. bacteriovorus are predicted to contribute to health, however, an overgrowth of Prevotella was observed due to exposure to M. aeruginosavorus. This study further addresses safety concerns associated with the potential use of predatory bacteria to treat infections.

With the rise of multi-drug resistant (MDR) bacterial infections over the last decade1, researchers have begun to develop new ways to treat infections for the future post-antibiotic era2. One alternative therapy currently being explored is the use of predatory bacteria3. Bdellovibrio bacteriovorus and Micavibrio aeruginosavorus are Gram-negative Delta- and Alphaproteobacteria, respectively, and are obligate predators of other Gram-negative bacteria4,5. B. bacteriovorus use a type IV pilus to attach to a prey cell, then invades across the cell membrane to establish itself in the periplasm of the prey forming a bdelloplast, before digesting the prey cell from within6,7,8. When the nutrients are exhausted, the growing B. bacteriovorus will divide, lyse the bdelloplast, and then continue to seek new prey. In contrast, M. aeruginosavorus are ‘vampire-like’ bacteria, as they attach to the prey cell outer membrane to leach nutrients from the outside5,9,10.

Predatory bacteria present several advantages that promote their potential use to combat bacterial infection. The effectiveness of both B. bacteriovorus and M. aeruginosavorus to limit the growth of key human pathogens, including those causing MDR infections11, in vitro is well documented3,12,13,14. Unlike with the use of traditional antibiotics, development of genetically-stable resistance to predation has yet to be confirmed15. Predatory bacteria are also obligate bacterial predators, shown not to harm many mammalian cell lines16,17,18. However, most studies focusing on predatory bacteria have been performed using in vitro methodologies, and thus the full effect of introducing predatory bacteria into a live host has yet to be determined.

The safety of administering predatory bacteria in vivo has already been examined in multiple animal models, such as mice, rats, rabbits, guinea pigs, and chicks3,19,20,21,22,23. It was recently demonstrated that intranasal or intravenous inoculation of high doses of either B. bacteriovorus or M. aeruginosavorus in mice resulted in no apparent pathological effects and induced a modest immune response that returned to physiological levels by 24 hours post-inoculation20. Additionally, both predators were not detected in animal tissues by 48 hours post-inoculation, potentially removed by innate immune mechanisms20. While the safety of introducing predatory bacteria into an animal host through respiratory and intravenous routes has already been assessed, we are interested in the effect predatory bacteria would have on the immune system within the gastrointestinal tract and on its residing commensal bacterial microbiota. This inquiry is of interest, not only because predatory bacteria have the capability to control bacterial populations, but also because evidence has suggested that B. bacteriovorus may actually contribute to health by being a member of the healthy human gut microbiota24.

As B. bacteriovorus might not survive within the acidic environment of the stomach19, administering predatory bacteria through oral gavage would not allow us to assess the effect they would have within the colon. To this end, intrarectal inoculations of rats with predatory bacteria were performed in this study to determine the effect on the gut immune response and the commensal bacterial microbiota populations. The work presented here further addresses some of the safety concerns associated with the use of predatory bacteria as a novel anti-microbial therapeutic.


Host Morbidity

To examine the effect of introducing predatory bacteria into the gut on host morbidity, we inoculated 8.2 × 108 plaque forming units (PFU)/rat of B. bacteriovorus 109J or 4.6 × 107 PFU/rat of M. aeruginosavorus ARL-13 into the colons of two groups of Sprague Dawley (SD) rats (64 rats each) via a rectal catheter (Fig. 1). Two more groups of rats were inoculated with the vehicle, phosphate buffered saline (PBS; 64 rats), or sub-lethal concentrations of 9.4 × 107 CFU/rat of Klebsiella pneumoniae ATCC 43826 (16 rats), a known pathogen. The aforementioned animals and treatment groups were used for all subsequent experiments described in this study. Rats were visually monitored for up to seven days for any signs of infection, illness or discomfort. At seven days post-inoculation, all 192 rats that were inoculated with PBS or with predatory bacteria over the course of the experiment were healthy and showed no adverse effects. Overall, our data indicate that introducing predatory bacteria into the gut have no effect on SD rat host morbidity.

Figure 1
Intrarectal inoculation of predatory bacteria into rats.

Histological Examination

To determine if predatory bacteria can cause detrimental histopathological effects within the colon, histological examination of colon tissue was performed on the previously described SD rats inoculated with PBS, B. bacteriovorus, M. aeruginosavorus, or K. pneumoniae, sacrificed at three, 24, 48 hours or seven days post-inoculation. Examination was performed by a pathologist blinded to each specimen’s treatment group, and findings were similar across all groups. Cross-sections of harvested colons were lined by a simple columnar epithelium and contained numerous evenly spaced glands (Fig. 2). Occasional mitotic figures were seen and the lamina propria and submucosa contained scattered lymphocytes and eosinophils (Fig. 2). Deposits of mucosa-associated lymphoid tissue were seen in some specimens (Fig. 2). Some specimens exhibited focal mucosal erosion, however, this was not accompanied by an observed inflammatory reaction and was most likely due to post-mortem mechanical abrasion. No other abnormalities were noted, indicating that introducing predatory bacteria into the rat gut results in no visually abnormal pathological effect.

Figure 2
Histological examination of distal colons from rats after intrarectal inoculation of predatory bacteria.

Host Immune Response

To examine the effects of predatory bacteria on the rat gut immune response, the previously described rat groups were sacrificed at three, 24, 48 hours and seven days post-inoculation (16 rats/time point for the PBS and predatory bacteria groups; eight rats/time point for K. pneumoniae) when distal colon samples were harvested for ELISA. At three, 24 and 48 hours, none of the rats exposed to any of the predatory bacteria strains exhibited significant increases in any inflammatory protein assayed (Fig. 3). A 6.8- and 9.6-fold increase in IL-13 in rats exposed to B. bacteriovorus and M. aeruginosavorus, respectively, was measured at seven days post-inoculation, however, none of the other inflammatory cytokines demonstrated substantial change compared to control at that time point (Fig. 3). Rats inoculated with K. pneumoniae demonstrated higher inductions of most inflammatory cytokines tested (particularly a 6.4-fold increase in IFNγ) by 24 hours post-inoculation (Fig. 3). In comparison, rats inoculated with B. bacteriovorus and M. aeruginosavorus demonstrated no change and a 10-fold decrease in IFNγ, respectively (Fig. 3). Collectively, our data indicate that B. bacteriovorus and M. aeruginosavorus do not provoke a robust inflammatory response within the gastrointestinal tract.

Figure 3
Inflammatory protein profile within rat colons in response to intrarectal inoculation of predatory bacteria.

Whole Bacterial Microbiota Overview

To examine the effect of predatory bacteria on the gut commensal bacterial microbiota, feces were collected daily for seven days from five rats from each treatment group. Sequencing of the V4 region of the 16S rRNA gene was performed on DNA extracted from feces in order to determine if there were significant changes in the bacterial communities in the gut. 16S rRNA gene sequencing produced 8,110,049 quality-filtered reads, giving a mean sample depth of 81,100 reads. Sequence reads per sample ranged from 23,302 to a maximum of 263,152 filtered reads.

At week’s end, linear regression analysis determined there was no significant difference in the observed or Shannon diversity (alpha diversity) for either B. bacteriovorus-treated or M. aeruginosavorus-treated samples compared to pre-treatment (Fig. 4). M. aeruginosavorus-treated samples had significantly less alpha diversity than both PBS and B. bacteriovorus-treated samples before treatment at day 0 (Fig. 4). Additionally, samples appeared to separate slightly according to day feces were collected in weighted ordination analysis, which compares community composition, of all samples sequenced (Fig. 5). PERMANOVA found time point (p < 0.002) and cage number which the animal was housed (p < 0.001) significantly contributed to bacterial microbiota beta diversity differences. Samples also clustered by treatment and sub-clustered according to cage number in hierarchical clustering analysis (Supp. Fig. S1).

Figure 4
Alpha-diversity estimates.
Figure 5
Weighted ordination (beta diversity).

Bacterial Microbiota Changes Over Time

The changes in gut bacterial populations due to each individual treatment over time were analyzed. In samples inoculated with the vehicle, PBS, genus Lactobacillus (and family Lactobacillaceae) relative abundance significantly decreased nearly 17-fold (p = 0.007), while families Ruminococcaceae and Peptostreptococcaceae relative abundance increased 1.4- (p = 0.011) and 3.7-fold (p = 0.002), respectively over time (Fig. 6A,B, Supp. Figs S2A and S3A). No significant shifts or concordant patterns were observed in the Gram-negative to Gram-positive bacterial ratios present in the microbiota (Fig. 6C, Supp. Fig. S4A). Bacterial populations slightly shifted according to the day feces were collected in weighted ordination analysis (p = 0.001; Supp. Fig. S5A) and samples clustered according to cage number in hierarchical clustering analysis (Supp. Fig. S6). There were no significantly differential operational taxonomic units (OTUs) detected out of 904 features tested. OTUs did not have a significantly large enough shift in relative abundance (absolute value of log2 fold change greater than one) over time.

Figure 6
Proportional taxa abundance over time.

In animals inoculated with B. bacteriovorus, genus Coprococcus significantly increased 2.2-fold (p = 0.01) in relative abundance (family Peptostreptococcaceae increased 2.7-fold, p = 0.01) and Lactobacillus significantly decreased 6.4-fold (p = 0.001) over the seven days (Lactobacillaceae increased 6.7-fold, p = 0.01; Fig. 6A,B, Supp. Figs S2B and S3B). While a mean 11% shift in the abundance from Gram-positive to Gram-negative bacteria was detected at day 7, there was high variation between individual rats (Fig. 6C, Supp. Fig. S4B). In an attempt to detect B. bacteriovorus in our sequencing reads, we detected four OTUs belonging to the Deltaproteobacteria class (Desulfovibrio, Bilophila, and two unclassified members of the Desulfovibrionaceae family and Myxococcales order). Additionally, the relative abundance of Deltaproteobacteria in the microbiota decreased on average nearly 4.0-fold at two days post-inoculation before increasing over the rest of the week. Principal coordinate analysis (PCoA) plots showed that bacterial populations appeared to shift according to the time feces were collected and formed two main clusters correlating with time point: one cluster was associated with the earlier time points of 0–2 days, and a second cluster with 3–7 days (p = 0.001; Supp. Fig. S5B). Samples further sub-clustered according to cage number according to hierarchical clustering analysis (Supp. Fig. S7). The bacterial communities in the rat gut were relatively stable after inoculation with predatory bacteria over time. Only two OTUs (out of 911 features tested), both belonging to Firmicutes, were significantly differentially abundant. Coprococcus eutactus had a 1.13 log2 fold increase per day (p < 0.05) and an unidentified member of Clostridiales had a 1.03 log2 fold decrease per day (p < 0.05).

In rats inoculated with M. aeruginosavorus, genus Akkermansia (family Verrucomicrobiaceae) relative abundance significantly decreased over 9.2-fold (p = 0.0002) and genus Prevotella (family Paraprevotellaceae) significantly increased 219-fold (p = 0.0004) over time (Fig. 6A,B, Supp. Figs S2C and S3C). As with the other treatments, Lactobacillus relative abundance also decreased 8.5-fold compared to pre-treatment (Fig. 6B). No significant shifts were detected in Gram-negative to Gram-positive bacterial ratios present in the microbiota (Fig. 6C, Supp. Fig. S4C). When attempting to detect M. aeruginosavorus in our sequencing reads, there did not appear to be a concordant trend in Alphaproteobacteria over time, however, 12 OTUs were identified as Alphaproteobacteria, including two which had 89% and 84% matches to M. aeruginosavorus ARL-13. Populations appeared to shift along the coordinate axes in weighted ordination analysis (p = 0.001; Supp. Fig. S5C and S8). Again, bacterial communities were relatively stable, as only one OTU (out of 910 features tested), from Prevotella, had a significant 1.09 log2 fold increase per day (p < 0.05).

Bacterial Microbiota Changes Due to Predatory Bacteria

The microbiota population changes due to inoculation with predatory bacteria compared to PBS control was determined next. Significantly, B. bacteriovorus-treated samples had 1.9-fold less Verrucomicrobiaceae than PBS samples (p < 0.001; Supp. Fig. S9A). In PCoA plots, B. bacteriovorus-treated samples clustered by time point and along both coordinate axes according to the day feces were collected (p = 0.001; Supp. Fig. S10A). Hierarchical clustering analysis showed samples also clustered according to cage number and treatment group (Supp. Fig. S11). There were a total of 166 differentially abundant OTUs (out of 911 features tested). Eighty-six OTUs, 77 of which were Firmicutes, were significantly overabundant in samples treated with B. bacteriovorus compared to PBS (p < 0.05; Fig. 7). Classified taxa at the genus level that were significantly overabundant in B. bacteriovorus-treated samples (n = OTU count) include Blautia (n = 6), Collinsella (n = 1), Coprococcus (n = 4), Lactococcus (n = 1), Odoribacter (n = 2), Oscollospira (n = 4), Roseburia (n = 1), and Ruminococcus (n = 1). An additional 69 OTUs identified as Firmicutes were also significantly reduced in B. bacteriovorus-treated samples (p < 0.05; Fig. 7). Genera that were found to be significantly reduced in B. bacteriovorus-treated samples compared to PBS include Adlercreutzia (n = 1), Akkermansia (n = 4), Anaeroplasma (n = 1), Coprococcus (n = 1), Dorea (n = 1), Oscillospira (n = 4), Ruminococcus (n = 3), Staphylococcus (n = 1), and Streptococcus (n = 1). Interestingly, OTUs with some of the highest log2 fold reduction were Gram-positive genera.

Figure 7
Differentially abundant features.

M. aeruginosavorus-treated samples had significantly higher relative abundances of Bacteroidaceae, Alcaligenaceae, and Paraprevotellacaeae than PBS-treated samples (p < 0.001; Supp. Fig. S9B). Significantly lower relative abundances of Ruminococcaceae, Lachnospiraceae and an unclassified family from Clostridiales were also observed in M. aeruginosavorus-treated samples compared to PBS (p < 0.001; Fig. 6B). M. aeruginosavorus-treated bacterial microbiotas clustered and shifted along the coordinate axes according to treatment in weighted ordination analysis (p = 0.001; Supp. Fig. S10B). Samples also clustered primarily by treatment, then by cage number, in hierarchical clustering analysis (Supp. Fig. S12). There were a total of 280 differentially abundant OTUs (out of 912 features tested), 122 of which were overabundant in M. aeruginosavorus-treated samples compared to PBS (p < 0.05; Fig. 7). Taxa at the genus level that were significantly overabundant in M. aeruginosavorus-treated samples include Akkermansia (n = 1), Anaerostipes (n = 1), Bacteroides (n = 7), Blautia (n = 4), Clostridium (n = 1), Coprococcus (n = 2), Dorea (n = 1), Lactobacillus (n = 3), Lactococcus (n = 1), Odoribacter (n = 2), Oscillospira (n = 2), Prevotella (n = 1), Ruminococcus (n = 2), Sutterella (n = 2), and Turicibacter (n = 1). 153 OTUs from Firmicutes were also found to be significantly reduced in M. aeruginosavorus-treated samples (p < 0.05; Fig. 7). Genera that were found to be significantly reduced in samples treated with M. aeruginosavorus compared to PBS include Adlercreutzia (n = 2), Akkermansia (n = 1), Bacteroides (n = 1), Coprococcus (n = 7), Dorea (n = 3), Lactobacillus (n = 1), Oscillospira (n = 6), Ruminococcus (n = 2), and Staphylococcus (n = 1). Once again, some of the OTUs that had the highest log2 fold reduction were from Gram-positive genera. Overall, the data show that there were minimal shifts in the bacterial communities over the course of the experiment, indicating that introducing predatory bacteria into the gastrointestinal tract of rats does not have a significant impact on the gut bacterial populations based on 16 S rRNA gene sequence analysis.


With the limited amount of new antibiotics being developed, scientists are looking for innovative ways to combat bacterial infection2, one of which is predatory bacteria. While the ability of predatory bacteria to control human pathogens in vitro is well documented, questions remain regarding the safety of introducing bacterial predators into a mammal. Here, our study focuses on the effect of administering predatory bacteria into the gastrointestinal tract of mammals.

We began by investigating the effect of introducing predatory bacteria into the gastrointestinal tract on rat morbidity, histopathology of the gut and its immune response. We intrarectally inoculated rats with PBS, predatory bacteria or K. pneumoniae and sacrificed animals at three, 24, 48 hours and seven days to harvest organs for analysis. The dose concentration of predatory bacteria administered was similar to our previous study examining the safety and efficacy of predatory bacteria within rat lungs23, however, the volume at which the dose was administered was increased from 50 μl to 250 μl in order to cover the larger surface area of the colon. K. pneumoniae is a Gram-negative opportunistic pathogen responsible for many hospital-acquired infections that can readily colonize the human gastrointestinal tract25, and while any Gram-negative bacterium could have been used as a comparison in this study, we sought to use a clinically relevant strain known to cause disease within the gut. A sub-lethal dose of K. pneumoniae was used in order to ensure rats survived the length of the experiment. All 128 rats inoculated with predatory bacteria into their gastrointestinal tracts were healthy with no visual adverse effects. Histology of the gut revealed no apparent inflammation or pathological effects due to predatory bacteria. Furthermore, none of the rats exposed to any of the predatory bacteria strains exhibited significant increases in any inflammatory protein assayed by 48 hours post-inoculation.

These results are consistent with previous studies that have demonstrated that predatory bacteria are non-toxic in different animal models and induce only a modest inflammatory response19,20,21,22. Interestingly, IL-13 levels remained elevated at day seven in rats inoculated with B. bacteriovorus or M. aeruginosavorus. Levels of IL-13 were slightly higher in rats inoculated with M. aeruginosavorus than with B. bacteriovorus at day seven, possibly explained by the fact M. aeruginosavorus are epibiotic predators and are potentially more easily accessible by the immune response elements. IL-13 has anti-inflammatory properties and is associated primarily with the induction of disease in the airways26,27,28, however, studies have shown IL-13 secreted by Th2 cells induce changes that are required to remove intracellular parasites from the mouse gut29,30,31. IL-13 can also regulate matrix metalloproteinases to protect against excessive inflammation in many tissues32, and it is possible that physiological changes in the gut due to predatory bacteria are being corrected through similar mechanisms. Future studies will focus on further understanding how the immune response specifically reacts to remove predatory bacteria from the gut environment.

We next looked to determine the effect of introducing predatory bacteria into the gut on the commensal microbiota over time. Samples inoculated with PBS showed slight shifts as time progressed. Specifically, a significant decrease in Lactobacillus, a Gram-positive genus, was seen across all treatment groups, including PBS, over time, signaling administration of PBS itself may have induced some minor microbiota changes. However, in samples inoculated with M. aeruginosavorus, this decrease in Lactobacillus was accompanied with a significant increase of Prevotella, which is a Gram-negative genus. An overgrowth of Prevotella and a reduction of Lactobacillus have been previously correlated with the onset of osteomyelitis in mice33. Interestingly, a previous study looking at the effect of B. bacteriovorus HD100 on multispecies oral communities found the periplasmic predator able to prey on Prevotella intermedia in co-cultures, but not in multispecies models34. In addition, a decrease in Akkermansia, a Gram-negative genus, was observed. Studies have shown Akkermansia to be present in higher abundances in healthy subjects compared to those with inflammatory bowel syndrome35, obesity and type 2 diabetes36. Although the measured changes in the bacterial microbiota might signal a potential negative outcome due to introducing M. aeruginosavorus into the gut, no changes in animal well-being or behavior was observed.

In samples inoculated with B. bacteriovorus, there was only one significant differential OTU which was an increase in Coprococcus eutactus, a Gram-positive genus. A previous study found decreased abundances of Coprococcus eutactus in inflammatory bowel syndrome (IBS) patients suffering with severe IBS-related symptoms37. B. bacteriovorus treatment may be able to increase Coprococcus eutactus abundances in these patients, potentially reversing microbiota dysbiosis. Overall, there were minimal shifts in other bacterial communities within each treatment over the seven days due to inoculation with predatory bacteria.

While we were unable conclusively to detect B. bacteriovorus in our fecal samples, we did detect four OTUs matching Deltaproteobacteria. In addition, a decrease in Deltaproteobacteria was seen at two days post-inoculation, however, populations significantly increased from that point forward for the rest of the week. One might suggest that the changes in Deltaproteobacteria relative abundance over the week could be the result of B. bacteriovorus predation early in the week before bacterial communities began to stabilize. A previous study was also not able to detect B. bacteriovorus in the fecal and cecal contents of young chicks after oral administration of predatory bacteria (and an antacid), signaling that exogenous B. bacteriovorus introduced into the gut environment may be short-lived19. However, it is important to note the relative abundance of Deltaproteobacteria is very low in the context of the whole bacterial microbiota community (assuming the rat gut contains ~1013 bacterial cells, the final ratio of predatory bacteria to commensal bacteria was approximately one in 100,000), and sequencing at a deeper depth will be required to fully understand these changes at a finer taxonomic level. Similar to our own results, the same chick study also observed an average ~10-fold decrease in Lactobacillus abundances in chick cecal contents after oral administration of B. bacteriovorus HD10019. We were able more conclusively to detect M. aeruginosavorus in our samples, as we found 12 OTUs matching to Alphaproteobacteria, two of which had high homology to M. aeruginosavorus.

Interestingly, there is evidence that predatory bacteria may already be present in healthy human flora. Multiple studies have detected B. bacteriovorus within the gut22,24,38,39 and oral40 microbiomes of humans and other animals. A recent study argued that further introducing bacterial predators back into the gut microbiota may alleviate some of the diseases associated with dysbiosis and loss of microbial diversity in the western world41. B. bacteriovorus populations are governed by the Lotka-Volterra prey-predator oscillation42, where increases in B. bacteriovorus abundances are balanced by phenotypic resistance to predation by the prey15. Therefore, it could be predicted that B. bacteriovorus present in the gastrointestinal tract may act as an ecological balancer of gut microbial populations contributing to health. In fact, a study found B. bacteriovorus are present in higher abundance within the gastrointestinal tract of healthy individuals compared to patients with Inflammatory Bowel Diseases, Celiac disease and Cystic fibrosis24.

It was hypothesized that since B. bacteriovorus and M. aeruginosavorus are obligate Gram-negative predators, they would shift the abundance between Gram-positive and Gram-negative bacteria over time. We did not observe any significant shift or concordant pattern in the Gram-negative to Gram-positive bacterial ratios present in the microbiota due to either B. bacteriovorus or M. aeruginosavorus treatment. We determined 166 and 280 differentially abundant OTUs between B. bacteriovorus or M. aeruginosavorus-treated samples, respectively, and PBS control. Interestingly, the majority of classified OTUs either significantly reduced or increased by inoculation with predatory bacteria belong to Firmicutes, a largely Gram-positive phylum. A possible explanation for this result is that predatory bacteria are not directly reducing Gram-positive bacteria, but rather microbiota changes in the gastrointestinal tract brought about due to the presence of predatory bacteria could allow certain bacterial populations to displace other populations, regardless of whether they are Gram-negative or Gram-positive genera. The previous study that orally administered predatory bacteria to young chicks found that while Lactobacillus populations in the gastrointestinal tract significantly decreased in abundance, Streptococcus populations increased in abundance19, signaling that changes in the gut microbiota cannot simply be predicted based on whether the bacterial prey of interest is Gram-negative or Gram-positive. Interestingly, in our own study, the only significantly differential OTU representing the Streptococcus genus was found to be reduced in samples treated with B. bacteriovorus. It is also possible that since Firmicutes are the dominant phyla in the gut bacterial microbiota, it is easier to detect changes in these populations due to sheer relative abundance compared to other phyla. Population changes in other phyla occupying the gastrointestinal tract may be detected if sequencing depth is increased in future studies.

Some of the other OTUs with the highest log2 fold increases due to either B. bacteriovorus or M. aeruginosavorus compared to PBS belonged to the genus Blautia. Blautia is a Gram-positive genus associated with normal health flora and has been shown to be one of the members of the gut microbiota to play an important role in recovery from Vibrio cholerae infection in mice43. Thus, it is possible that the overabundance in Blautia is in direct response to the introduction of Bdellovibrio and Micavibrio into the gut environment. Among genera that contained OTUs that were both increased and decreased due to predatory bacteria exposure were Coprococcus and Oscillospira. Oscillospira is a Gram-positive genus that has never been cultured in the laboratory, but is associated with leanness and health44. While it is unlikely that these two Gram-positive genera are being preyed upon, differential rates of predation by B. bacteriovorus and M. aeruginosavorus on species within the same genus have been observed before11,12. Thus, it is not surprising to see many genera belonging to the same taxonomic family being affected differently due to predatory bacteria inoculation.

It is important to view the changes in the bacterial microbiota due to inoculation of predatory bacteria in the context of population shifts brought about by antibiotics. Antibiotics, while an efficient treatment for infection, have off-target effects that can cause dysbiosis within the gut microbiome, often leading to significant decreases in bacterial diversity45,46,47,48,49. In fact, a study demonstrated that changes in the gut microbiota due to antibiotic use may last for up to a year before returning to previous abundances50. Furthermore, the same study demonstrated that exposure to different antibiotics enriched genes associated with antibiotic resistance50. In contrast, B. bacteriovorus minimally shifted the bacterial microbiota as compared to antibiotics, and with research beginning to demonstrate a link between health and normal gastrointestinal flora, this represents a major advantage if predatory bacteria are to be developed into a new treatment for infection.

In conclusion, our results suggest that introducing predatory bacteria into the rat gut does not result in any measurable adverse pathological effects and does not cause a substantial immune response within the gut. Furthermore, the limited changes in gut bacterial populations due to intrarectal inoculation of B. bacteriovorus directly into the gastrointestinal tract were mostly associated with healthy benefits, while potential adverse negative effects were seen in microbiota changes due to M. aeruginosavorus. One could suggest that increasing the number of doses and the concentration of predatory bacteria may result in more substantial changes in the bacterial microbiota, and future studies will address this. The limited population shifts in the microbiota due to predatory bacteria compares favorably against the off-target effects of antibiotics. With the increased prevalence of antibiotic resistant infections, this study provides further support for developing predatory bacteria into a novel antimicrobial treatment. Future studies will focus on further understanding how predatory bacteria are removed from the gut by the immune response, as well as increasing sequencing depth to further decipher changes in the microbiota at the species level due to administration of predatory bacteria.


Bacterial strains and growth conditions

Bacterial strains used in this study were Bdellovibrio bacteriovorus 109 J (ATCC 43826), Micavibrio aeruginosavorus strain ARL-1310, and Klebsiella pneumoniae ATCC 43816. K. pneumoniae was cultured in LB media. Predatory bacteria were cultured and processed as previously described17. In brief, predator stock lysates were made by co-culturing the predators with Escherichia coli WM3064 used as prey, and incubating at 30 °C on a rotary shaker for 24 (B. bacteriovorus) and 72 hours (M. aeruginosavorus) until the cultures cleared. Fresh predator cultures were prepared from the stock lysates to obtain higher concentration of predators for the experiments, as described previously17. The fresh co-cultures were filtered twice through a 0.45-μm Millex pore-size filter (Millipore) and further purified and concentrated by three sequential centrifugations at 29,000 g for 45 minutes using a Sorvall LYNX 4000 centrifuge (Thermo Fisher Scientific Inc). The pellet was washed and re-suspended in 50 mL of PBS after each cycle. A plaque-forming unit (PFU) value between ~5 × 109 to 5 × 1010 PFU/mL for B. bacteriovorus and ~5 × 108 to 5 × 109 PFU/mL for M. aeruginosavorus, was obtained when the final predator pellet was re-suspended in 1–2 mL of PBS solution to reach a final optical density (OD600) of 0.2 ± 0.02 for B. bacteriovorus and 0.1 ± 0.02 for M. aeruginosavorus. Fifty μl aliquots of predator samples were plated on DAP-supplemented LB agar, nutrient agar and TSB-blood plates to ensure that the samples were free from contamination and prey cells.


Wild-type 4–6 week old male Sprague Dawley (SD) rats were purchased from Charles River Laboratories (Wilmington, MA, USA). A maximum of four rats were housed per cage. All rats were housed under pathogen-free conditions at the Rutgers New Jersey Medical School animal facility. All experiments were performed in accordance with the protocols approved by the Institutional Animal Care and Use Committee of Rutgers New Jersey Medical School (protocol #15012) and the Animal Care and Use Review Office of the U.S. Army Medical Research and Material Command were followed in handling the animals.

Intrarectal Inoculation

Predatory bacteria were introduced by intracolonic instillation51 of Sprague Dawley (SD) rats. Animals were anaesthetized with 4% isoflurane in oxygen for five minutes. A 3Fr polyurethane catheter (Access Technologies, Skokie, Illinois, USA) with a 1 mL syringe attached via a Luer Lock was inserted intrarectally to approximately 8 cm proximal to the anus, at which 250 μl of bacterial suspension (diluted in PBS) was slowly injected into the colon. Rats were introduced with either PBS (vehicle), B. bacteriovorus 109J, M. aeruginosavorus ARL-13, or K. pneumoniae ATCC 43826. To prevent leakage of the bacterial suspension, animals were held at an angle with anus pinched closed for 30 seconds after removing the catheter from the colon. A smaller number of rats were used in the K. pneumoniae group in order to limit the number of animals being sacrificed. Being mindful of animal wellbeing, we also did not keep any rats inoculated with K. pneumoniae past 24 hours post-inoculation. After initial inoculation, animals were observed and visually assessed for signs of infection, illness and discomfort. To assess the immune response, rats were sacrificed at three, 24, 48 hrs and 7 days post-exposure, when colon samples were collected. Samples were used for histological examination and host immune response profiling (ELISA).

Histological Examination

Distal colon samples were stored in formalin at 4 °C before examination. All histopathological examination was performed by a pathologist blinded to each specimen’s treatment group. Formalin-fixed organ segments from infected rats were paraffin-embedded and stained with hematoxylin and eosin (H&E) for cellular composition as previously described. Stained sections were analyzed and photographed using an EVOS FL Cell Imaging System (Life Technologies, Carlsbad, CA).

Host Immune Response Profiling (ELISA)

Samples were prepared as previously described20. Cytokines were measured using a V-Plex Proinflammatory Panel2 (rat) Kit (K15059D-1); Meso Scale Discovery) according to manufacturer’s instructions, and read on a SECTOR Imager 2400 (Meso Scale Discovery).

Fecal Collection and DNA Extraction

To determine the effect of predatory bacteria on the commensal gut bacterial microbiota, feces were collected from five rats from each treatment group for each day for up to seven days post-instillation and frozen immediately at −80 °C. DNA from feces was extracted using the PowerFecal DNA Isolation Kit (Mo Bio Laboratories). DNA quantity and quality was assessed using an Agilent 2100 Bioanalyzer (Santa Clara, CA, USA).

Library Preparation

Purified DNA was sent to Second Genome Solutions (South San Francisco, CA, USA) for library preparation and sequencing. On arrival, all samples were quantified via the Qubit® Quant-iT dsDNA High Sensitivity Kit (Invitrogen, Life Technologies, Grand Island, NY) to ensure that they met minimum concentration and mass of DNA. To enrich the sample for bacterial 16S V4 rRNA genes, DNA was amplified utilizing fusion primers designed against the surrounding conserved regions which are tailed with sequences to incorporate Illumina (San Diego, CA, USA) adapters and indexing barcodes. Each sample was PCR amplified with two differently bar-coded V4 fusion primers. Samples that met post-PCR quantification minimums were advanced for pooling and sequencing. For each sample, amplified products were concentrated using a solid-phase reversible immobilization method for the purification of PCR products and quantified by qPCR.

Commensal Microbiota Profiling

A pool containing 16S V4 enriched, bar-coded samples were loaded into a MiSeq® reagent cartridge, and then onto the instrument along with the flow cell. After cluster formation on the MiSeq instrument, the amplicons were sequenced for 250 cycles with custom primers designed for paired-end sequencing. Sequenced paired-end reads were merged, quality filtered, and de-replicated with USEARCH52. Resulting unique sequences were then clustered at 97% similarity by UPARSE and a representative consensus sequence per de novo OTU was determined. Sequences that passed quality filtering were then mapped to a set of representative consensus sequences to generate an OTU abundance table. Representative OTU sequences were assigned taxonomic classification via Mothur’s bayesian classifier at 80% confidence; the classifier was trained against the Greengenes53 reference database of 16S rRNA gene sequences clustered at 99%.

Statistical Analysis

For ELISA, significant differences between treatment groups and respective control were analyzed using ANOVA. Alpha-diversity differences between beginning and end of treatment were compared using linear regression, controlling for animal ID. PERMANOVA54 (paired by animal ID) using distance matrices and was performed for each variable of interest to determine if they significantly contributed to the beta-diversity of the samples. Abundance-weighted sample pair-wise differences were calculated using Bray-Curtis55 dissimilarity and plotted using the Principal Coordinate Analysis (PCoA). Hierarchical clustering was done using the Ward 2 method. Differences in mean population abundances over time in the microbiota were compared using linear regression, controlling for animal ID. Univariate differential abundance of OTUs was tested using the DESeq2 package56. DESeq2 (paired by animal ID) was run under default settings and q-values were calculated with the Benjamin-Hochberg57 procedure to correct p-values, controlling for false discovery rates (FDR). OTUs were considered significant if their FDR-corrected p-value was less than or equal to 0.05, and the absolute value of the log2 fold change was greater than or equal to 1.

Additional Information

How to cite this article: Shatzkes, K. et al. Effect of predatory bacteria on the gut bacterial microbiota in rats. Sci. Rep. 7, 43483; doi: 10.1038/srep43483 (2017).

Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Material

Supplementary Information:


We are grateful to Juli-Anne Royes Russo (Connell Lab, Rutgers New Jersey Medical School) for administrative help in managing different aspects of the study. We thank Dr. Marcus B. Jones (Human Longevity, Inc.) and Erik Snesrud (Walter Reed Army Institute of Research) for reviewing the work and for their constructive comments. Research was sponsored by the U.S. Army Research Office and the Defense Advanced Research Projects Agency and was accomplished under Cooperative Agreement Number W911NF-15–2–0036. The views and conclusions contained in this document are those of the authors and should not be interpreted as representing the official policies, either expressed or implied, of the Army Research Office, DARPA, or the U.S. Government. The U.S. Government is authorized to reproduce and distribute reprints for Government purposes notwithstanding any copyright notation hereon.


The authors declare no competing financial interests.

Author Contributions K.S., N.D.C. and D.E.K conceived and designed the experiments. K.S., C.T., E.S., S.S., M.Z., S.G., and S.D. performed the experiments. K.S., C.T, and J.R. analyzed the data. K.S. drafted the manuscript. K.S., N.D.C. and D.E.K. supervised the study and revised the manuscript. All authors reviewed the manuscript.


  • WHO. Antimicrobial resistance: global report on surveillance 2014. (2014).
  • Czaplewski L. et al. . Alternatives to antibiotics-a pipeline portfolio review. Lancet Infect. Dis. 16, 239–251, doi: (2016).10.1016/s1473-3099(15)00466-1 [PubMed] [Cross Ref]
  • Dwidar M., Monnappa A. K. & Mitchell R. J. The dual probiotic and antibiotic nature of Bdellovibrio bacteriovorus. BMB Rep 45, 71–78, doi: (2012).10.5483/BMBRep.2012.45.2.71 [PubMed] [Cross Ref]
  • Stolp H. & Starr M. P. Bdellovibrio bacteriovorus gen. et sp. n., a predatory, ectoparasitic, and bacteriolytic microorganism. Antonie Van Leeuwenhoek 29, 217–248 (1963). [PubMed]
  • Lambina V. A., Afinogenova A. V., Romay Penobad Z., Konovalova S. M. & Andreev L. V. [New species of exoparasitic bacteria of the genus Micavibrio infecting gram-positive bacteria]. Mikrobiologiia 52, 777–780 (1983). [PubMed]
  • Chanyi R. M. & Koval S. F. Role of type IV pili in predation by Bdellovibrio bacteriovorus. PLoS One 9, e113404, doi: (2014).10.1371/journal.pone.0113404 [PMC free article] [PubMed] [Cross Ref]
  • Evans K. J., Lambert C. & Sockett R. E. Predation by Bdellovibrio bacteriovorus HD100 requires type IV pili. J. Bacteriol. 189, 4850–4859, doi: (2007).10.1128/jb.01942-06 [PMC free article] [PubMed] [Cross Ref]
  • Mahmoud K. K. & Koval S. F. Characterization of type IV pili in the life cycle of the predator bacterium Bdellovibrio. Microbiology 156, 1040–1051, doi: (2010).10.1099/mic.0.036137-0 [PubMed] [Cross Ref]
  • Lambina V. A., Afinogenova A. V., Romai Penabad S., Konovalova S. M. & Pushkareva A. P. [Micavibrio admirandus gen. et sp. nov]. Mikrobiologiia 51, 114–117 (1982). [PubMed]
  • Wang Z., Kadouri D. E. & Wu M. Genomic insights into an obligate epibiotic bacterial predator: Micavibrio aeruginosavorus ARL-13. BMC Genomics 12, 453, doi: (2011).10.1186/1471-2164-12-453 [PMC free article] [PubMed] [Cross Ref]
  • Kadouri D. E., To K., Shanks R. M. & Doi Y. Predatory bacteria: a potential ally against multidrug-resistant Gram-negative pathogens. PLoS One 8, e63397, doi: (2013).10.1371/journal.pone.0063397 [PMC free article] [PubMed] [Cross Ref]
  • Dashiff A., Junka R. A., Libera M. & Kadouri D. E. Predation of human pathogens by the predatory bacteria Micavibrio aeruginosavorus and Bdellovibrio bacteriovorus. J. Appl. Microbiol. 110, 431–444, doi: (2011).10.1111/j.1365-2672.2010.04900.x [PubMed] [Cross Ref]
  • Kadouri D. & O’Toole G. A. Susceptibility of biofilms to Bdellovibrio bacteriovorus attack. Appl. Environ. Microbiol. 71, 4044–4051, doi: (2005).10.1128/aem.71.7.4044-4051.2005 [PMC free article] [PubMed] [Cross Ref]
  • Kadouri D., Venzon N. C. & O’Toole G. A. Vulnerability of pathogenic biofilms to Micavibrio aeruginosavorus. Appl. Environ. Microbiol. 73, 605–614, doi: (2007).10.1128/aem.01893-06 [PMC free article] [PubMed] [Cross Ref]
  • Shemesh Y. & Jurkevitch E. Plastic phenotypic resistance to predation by Bdellovibrio and like organisms in bacterial prey. Environ. Microbiol. 6, 12–18 (2004). [PubMed]
  • Shanks R. M. et al. . An eye to a kill: using predatory bacteria to control Gram-negative pathogens associated with ocular infections. PLoS One 8, e66723, doi: (2013).10.1371/journal.pone.0066723 [PMC free article] [PubMed] [Cross Ref]
  • Gupta S., Tang C., Tran M. & Kadouri D. E. Effect of predatory bacteria on human cell lines. PLoS One 11, e0161242, doi: (2016).10.1371/journal.pone.0161242 [PMC free article] [PubMed] [Cross Ref]
  • Romanowski E. G. et al. . Predatory bacteria are nontoxic to the rabbit ocular surface. Sci Rep 6, 30987, doi: (2016).10.1038/srep30987 [PMC free article] [PubMed] [Cross Ref]
  • Atterbury R. J. et al. . Effects of orally administered Bdellovibrio bacteriovorus on the well-being and Salmonella colonization of young chicks. Appl. Environ. Microbiol. 77, 5794–5803, doi: (2011).10.1128/aem.00426-11 [PMC free article] [PubMed] [Cross Ref]
  • Shatzkes K. et al. . Examining the safety of respiratory and intravenous inoculation of Bdellovibrio bacteriovorus and Micavibrio aeruginosavorus in a mouse model. Sci. Rep. 5, 12899, doi: (2015).10.1038/srep12899 [PMC free article] [PubMed] [Cross Ref]
  • Verklova Z. S. [Study of the virulence, toxicity and immunogenicity of different strains of Bdellovibrio bacteriovorus]. Gig. Sanit. 38, 10–13 (1973). [PubMed]
  • Westergaard J. M. & Kramer T. T. Bdellovibrio and the intestinal flora of vertebrates. Appl. Environ. Microbiol. 34, 506–511 (1977). [PMC free article] [PubMed]
  • Shatzkes K. et al. . Predatory bacteria attenuate Klebsiella pneumoniae burden in rat lungs. mBio 7, doi: (2016).10.1128/mBio.01847-16 [PMC free article] [PubMed] [Cross Ref]
  • Iebba V. et al. . Higher prevalence and abundance of Bdellovibrio bacteriovorus in the human gut of healthy subjects. PLoS One 8, e61608, doi: (2013).10.1371/journal.pone.0061608 [PMC free article] [PubMed] [Cross Ref]
  • Maroncle N., Balestrino D., Rich C. & Forestier C. Identification of Klebsiella pneumoniae genes involved in intestinal colonization and adhesion using signature-tagged mutagenesis. Infect. Immun. 70, 4729–4734 (2002). [PMC free article] [PubMed]
  • Wynn T. A. IL-13 effector functions. Annu. Rev. Immunol. 21, 425–456, doi: (2003).10.1146/annurev.immunol.21.120601.141142 [PubMed] [Cross Ref]
  • Gour N. & Wills-Karp M. IL-4 and IL-13 signaling in allergic airway disease. Cytokine 75, 68–78, doi: (2015).10.1016/j.cyto.2015.05.014 [PMC free article] [PubMed] [Cross Ref]
  • Townley R. G., Sapkota M. & Sapkota K. IL-13 and its genetic variants: effect on current asthma treatments. Discov. Med. 12, 513–523 (2011). [PubMed]
  • Grencis R. K. & Bancroft A. J. Interleukin-13: a key mediator in resistance to gastrointestinal-dwelling nematode parasites. Clin. Rev. Allergy Immunol. 26, 51–60, doi: (2004).10.1385/criai:26:1:51 [PubMed] [Cross Ref]
  • Sun R. et al. . Interleukin-13 Receptor alpha1-dependent responses in the intestine are critical to parasite clearance. Infect. Immun. 84, 1032–1044, doi: (2016).10.1128/iai.00990-15 [PMC free article] [PubMed] [Cross Ref]
  • Mannon P. & Reinisch W. Interleukin 13 and its role in gut defence and inflammation. Gut 61, 1765–1773, doi: (2012).10.1136/gutjnl-2012-303461 [PubMed] [Cross Ref]
  • Bailey J. R. et al. . IL-13 promotes collagen accumulation in Crohn’s disease fibrosis by down-regulation of fibroblast MMP synthesis: a role for innate lymphoid cells? PLoS One 7, e52332, doi: (2012).10.1371/journal.pone.0052332 [PMC free article] [PubMed] [Cross Ref]
  • Lukens J. R. et al. . Dietary modulation of the microbiome affects autoinflammatory disease. Nature 516, 246–249, doi: (2014).10.1038/nature13788 [PMC free article] [PubMed] [Cross Ref]
  • Loozen G. et al. . Effect of Bdellovibrio bacteriovorus HD100 on multispecies oral communities. Anaerobe 35, 45–53, doi: (2015).10.1016/j.anaerobe.2014.09.011 [PubMed] [Cross Ref]
  • van Passel M. W. et al. . The genome of Akkermansia muciniphila, a dedicated intestinal mucin degrader, and its use in exploring intestinal metagenomes. PLoS One 6, e16876, doi: (2011).10.1371/journal.pone.0016876 [PMC free article] [PubMed] [Cross Ref]
  • Dao M. C. et al. . Akkermansia muciniphila and improved metabolic health during a dietary intervention in obesity: relationship with gut microbiome richness and ecology. Gut 65, 426–436, doi: (2016).10.1136/gutjnl-2014-308778 [PubMed] [Cross Ref]
  • Malinen E. et al. . Association of symptoms with gastrointestinal microbiota in irritable bowel syndrome. World J Gastroenterol 16, 4532–4540 (2010). [PMC free article] [PubMed]
  • Chu W. H. & Zhu W. Isolation of Bdellovibrio as biological therapeutic agents used for the treatment of Aeromonas hydrophila infection in fish. Zoonoses and public health 57, 258–264, doi: (2010).10.1111/j.1863-2378.2008.01224.x [PubMed] [Cross Ref]
  • Schwudke D., Strauch E., Krueger M. & Appel B. Taxonomic studies of predatory Bdellovibrios based on 16S rRNA analysis, ribotyping and the hit locus and characterization of isolates from the gut of animals. Syst. Appl. Microbiol. 24, 385–394, doi: (2001).10.1078/0723-2020-00042 [PubMed] [Cross Ref]
  • Dewhirst F. E. et al. . The human oral microbiome. J. Bacteriol. 192, 5002–5017, doi: (2010).10.1128/jb.00542-10 [PMC free article] [PubMed] [Cross Ref]
  • Mosca A., Leclerc M. & Hugot J. P. Gut microbiota diversity and human diseases: should we reintroduce key predators in our ecosystem? Front. Microbiol. 7, 455, doi: (2016).10.3389/fmicb.2016.00455 [PMC free article] [PubMed] [Cross Ref]
  • Varon M. & Zeigler B. P. Bacterial predator-prey interaction at low prey density. Appl. Environ. Microbiol. 36, 11–17 (1978). [PMC free article] [PubMed]
  • Hsiao A. et al. . Members of the human gut microbiota involved in recovery from Vibrio cholerae infection. Nature 515, 423–426, doi: (2014).10.1038/nature13738 [PMC free article] [PubMed] [Cross Ref]
  • Konikoff T. & Gophna U. Oscillospira: a central, enigmatic component of the human gut microbiota. Trends Microbiol 24, 523–524, doi: (2016).10.1016/j.tim.2016.02.015 [PubMed] [Cross Ref]
  • Greenwood C. et al. . Early empiric antibiotic use in preterm infants is associated with lower bacterial diversity and higher relative abundance of Enterobacter. J. Pediatr. 165, 23–29, doi: (2014).10.1016/j.jpeds.2014.01.010 [PMC free article] [PubMed] [Cross Ref]
  • Dethlefsen L., Huse S., Sogin M. L. & Relman D. A. The pervasive effects of an antibiotic on the human gut microbiota, as revealed by deep 16S rRNA sequencing. PLoS Biol. 6, e280, doi: (2008).10.1371/journal.pbio.0060280 [PMC free article] [PubMed] [Cross Ref]
  • Dethlefsen L. & Relman D. A. Incomplete recovery and individualized responses of the human distal gut microbiota to repeated antibiotic perturbation. Proc. Natl. Acad. Sci. USA 108 Suppl 1, 4554–4561, doi: (2011).10.1073/pnas.1000087107 [PubMed] [Cross Ref]
  • Lichtman J. S. et al. . Host-microbiota interactions in the pathogenesis of antibiotic-associated diseases. Cell reports 14, 1049–1061, doi: (2016).10.1016/j.celrep.2016.01.009 [PubMed] [Cross Ref]
  • Langdon A., Crook N. & Dantas G. The effects of antibiotics on the microbiome throughout development and alternative approaches for therapeutic modulation. Genome Med. 8, 1–16, doi: (2016).10.1186/s13073-016-0294-z [PMC free article] [PubMed] [Cross Ref]
  • Zaura E. et al. . Same exposure but two radically different responses to antibiotics: resilience of the salivary microbiome versus long-term microbial shifts in feces. mBio 6, e01693–01615, doi: (2015).10.1128/mBio.01693-15 [PMC free article] [PubMed] [Cross Ref]
  • Morampudi V. et al. . DNBS/TNBS colitis models: providing insights into inflammatory bowel disease and effects of dietary fat. e51297, doi: 10.3791/51297 (2014). [PubMed]
  • Edgar R. C. UPARSE: highly accurate OTU sequences from microbial amplicon reads. Nature methods 10, 996–998, doi: (2013).10.1038/nmeth.2604 [PubMed] [Cross Ref]
  • McDonald D. et al. . An improved Greengenes taxonomy with explicit ranks for ecological and evolutionary analyses of bacteria and archaea. The ISME journal 6, 610–618, doi: (2012).10.1038/ismej.2011.139 [PMC free article] [PubMed] [Cross Ref]
  • Anderson M. J. A new method for non-parametric multivariate analysis of variance. Austral Ecol. 26, 32–46, doi: (2001).10.1111/j.1442-9993.2001.01070.pp.x [Cross Ref]
  • Bray J. R. & Curtis J. T. An ordination of the upland forest communities of southern Wisconsin. Ecol. Monogr. 27, 325–349, doi: (1957).10.2307/1942268 [Cross Ref]
  • Love M. I., Huber W. & Anders S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol. 15, 550, doi: (2014).10.1186/s13059-014-0550-8 [PMC free article] [PubMed] [Cross Ref]
  • Benjamini Y. & Hochberg Y. Controlling the false discovery rate: a practical and powerful approach to multiple testing. Journal of the Royal Statistical Society. Series B (Methodological) 57, 289–300, doi: (1995).10.2307/2346101 [Cross Ref]

Articles from Scientific Reports are provided here courtesy of Nature Publishing Group