Utilizing the fecal microbiota to understand foal gut transitions from birth to weaning

Abstract

A healthy gastrointestinal (GI) tract with a properly established microbiota is necessary for a foal to develop into a healthy weanling. A foal’s health can be critically impacted by aberrations in the microbiome such as with diarrhea which can cause great morbidity and mortality in foals. In this study, we hypothesized that gut establishment in the foal transitioning from a diet of milk to a diet of grain, forage, and pasture would be detectable through analyses of the fecal microbiotas. Fecal samples from 37 sets of foals and mares were collected at multiple time points ranging from birth to weaning. Bacterial DNA was isolated from the samples, and the V4 domain of bacterial 16S rRNA genes were amplified via polymerase chain reaction. Next generation sequencing was then performed on the resulting amplicons, and analyses were performed to characterize the microbiome as well as the relative abundance of microbiota present. We found that bacterial population compositions followed a pattern throughout the early life of the foal in an age-dependent manner. As foals transitioned from milk consumption to a forage and grain diet, there were recognizable changes in fecal microbial compositions from initial populations predominant in the ability to metabolize milk to populations capable of utilizing fibrous plant material. We were also able to recognize differences in microbial populations amongst diarrheic foals as well as microbial population differences associated with differences in management styles between facilities. Future efforts will gauge the effects of lesser abundant bacterial populations that could also be essential to GI health, as well as to determine how associations between microbial population profiles and animal management practices can be used to inform strategies for improving upon the health and growth of horses overall.

About the Authors:

Ubaldo De La Torre

Roles Data curation, Formal analysis, Funding acquisition, Methodology, Writing – original draft, Writing – review & editing

Affiliation: Department of Animal Science, University of California Davis, Davis, California, United States of America

John D. Henderson

Roles Data curation, Methodology, Supervision, Writing – review & editing

Affiliation: Department of Animal Science, University of California Davis, Davis, California, United States of America

Kathleen L. Furtado

Roles Methodology, Project administration, Writing – review & editing

Affiliation: Department of Animal Science, University of California Davis, Davis, California, United States of America

Madeleine Pedroja

Roles Data curation, Methodology, Writing – review & editing

Affiliation: Department of Animal Science, University of California Davis, Davis, California, United States of America

O’Malley Elenamarie

Roles Data curation, Methodology

Affiliation: Department of Animal Science, University of California Davis, Davis, California, United States of America

Anthony Mora

Roles Methodology

Affiliation: Department of Animal Science, University of California Davis, Davis, California, United States of America

Monica Y. Pechanec

Roles Data curation, Methodology, Project administration, Writing – review & editing

Affiliation: Department of Animal Science, University of California Davis, Davis, California, United States of America

Elizabeth A. Maga

Roles Conceptualization, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Supervision, Writing – review & editing

Affiliation: Department of Animal Science, University of California Davis, Davis, California, United States of America

Michael J. Mienaltowski

Roles Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Resources, Software, Supervision, Validation, Visualization, Writing – original draft, Writing – review & editing

* E-mail: [email protected]

Affiliation: Department of Animal Science, University of California Davis, Davis, California, United States of America

ORCID logo http://orcid.org/0000-0001-6456-9894

Introduction

A foal grows from about 10% of its mature body weight at birth to as much as 50% of that weight by the time of weaning [1]. As the foal grows, synchronization occurs between the changes in dietary needs, changes in type of food consumed (e.g., changes in mare’s milk composition, introduction of creep feeding to transition to a solid diet), and shifts in the gut microbiota to bacterial populations that can more efficiently utilize the diet provided. Due to fairly recent advancements in ‘omic’ technologies, the importance of the microbiota on health is being realized because of the sudden increase of available information on gut microbiota composition and functions. The gut microbiota may even be seen as an organ system in the host given the important roles it plays in processing ingested organic matter [2]. The health of the host, or in this case the foal, is dependent upon these microbes and can be impacted by perturbances to the microbiota such as those caused by infectious diseases or antibiotic treatment.

Naturally, the mare provides some defense to the foal via immunoglobulins in colostrum and milk and levels of innate anti-microbial molecules like lysozyme in the milk [3]. These initial contacts with the mare may begin to provide the foal with early colonizing microbes. Studies have begun to understand which microbial populations comprise healthy and unhealthy gut microbiomes and how they may change once the foal no longer relies on the mare for food [4–7]. Since there are many risks to the well-being of a neonate’s GI health, developing methods to track and assess GI health would be advantageous. Bacterial community structures in the foal have been monitored using ribosomal intergenic spacer analysis which is a “fingerprint” of gut microbiota diversity but does not specifically delineate composition details [8]. Others have tried to specifically identify populations by culturing specific species from foals in the first 5 weeks of life, but only those populations capable of culture were studied [9]. Infectious bacterial populations have been specifically interrogated via development of microbial species-specific diagnostic tools [10–12].

This study utilizes Next-Generation Sequencing (NGS) technology to continue comprehensively analyzing gut microbial composition and establishment in foals as represented by fecal samples. We hypothesized that differences in gut establishment by age and by diarrhea status would be detectable in analyses of the fecal microbiotas of foals from birth to weaning. Here we report on identity and quantity of microbial species to understand the colonization of the gut microbiota in foals leading to the establishment of a reference catalog of bacteria present in the feces of foals that are either healthy or ill. The study affirms that NGS can be used as a tool to track and predict the GI health status of a foal from birth to weaning. Improved understanding of foal gut microbiota in various health states will aid in optimizing care of foals from postnatal to weaning stages.

Materials and methods

Horses and sample collection

The UC Davis IACUC approved our study (#18998): “Characterization of the Gut Microbiota of Foals from Birth to Weaning.” Voluntarily-voided fecal samples were collected from 37 sets of foals and mares from several breeds at three different farm locations (S1 Table). A majority of horses were either Thoroughbreds or American Quarter Horses. Management for the farms had similarities and differences; mares were fed combinations of supplemental grain and hay with consideration for pasture access at two farms (S2 Table) to provide nutrient requirements as recommended by the Committee on Nutrient Requirements of Horses (National Research Council Board on Agriculture and Natural Resources, National Academy of Sciences) [13, 14]. Samples were collected from foals at days 1, 7, 28, 60, and at weaning; samples were collected from mares at day 1 of the birth of the foals (S1 Table). The fecal samples were stored in polypropylene tubes at -20°C until DNA isolation was performed at about 30 days post-collection.

Bacterial DNA isolation, library preparation, and sequencing

One hundred sixty-three fecal samples were thawed and PCR-quality DNA was isolated based upon manufacturer’s instructions (ZR Fecal DNA Kit, Zymo Research) [15] with adaptations to increase centrifugation speeds. DNA concentration was determined using a NanoDrop UV spectrophotometer (ThermoFisher Scientific). The V4 domain of bacterial 16S rRNA genes was then amplified to generate libraries bar-coded by sample. Primers F515 (forward: 5’-GTGCCAGCMGCCGCGGTAA-3’) and R806 (5’-GGACTACHVGGGTWTCTAAT-3’) were used to amplify the V4 domain and included a unique barcode on each forward primer [16]. PCR was performed in triplicates for 25-ul reactions using GoTaq 2X Green Master Mix (ProMega) and programmed to follow: 1 initial step at 94°C for 3 min, followed by 35 cycles of 94°C for 45 sec, 50°C for 1 min, and 72°C for 90 sec, ending with a final extension at 72°C for 10 min. PCR amplification success was examined using agarose gel electrophoresis after triplicates were combined for each sample. Purification of PCR products was performed using QIAGEN’s PCR Purification Kit. Combined barcoded libraries were submitted to the University of California Davis Genome Center DNA Technologies Core for 250bp paired-end sequencing using the Illumina MiSeq platform. Raw sequence data are freely available at the Sequence Read Archive (SRA): Bio Project PRJNA475435, BioSample Accession SAMN09389119.

Microbiota analyses

DNA sequences from NGS were initially processed using QIIME 1.9.1 (Quantitative Insights Into Microbial Ecology) open-source software [17]. All sequences were demultiplexed, filtered, and assigned an operational taxonomic unit (OTU) using the Greengenes database (v13.8) at 97% identity in order to perform further diversity analyses. To filter sequences, a minimal fraction threshold of 0.005% of reads was used in QIIME. After filtering, samples were rarified to 3,590 sequences from each sample based on the samples with the lowest number of reads, and thus allowing for the incorporation of all samples in downstream analyses. QIIME was used to create rarefaction curves, taxonomical bar plots, and PCA (Principal Component Analysis) plots. OTU tables produced by QIIME were extracted and inputted into LEfSe (LDA Effect Size) and STAMP (Statistical Analysis of Metagenomic Profiles) software [18, 19] with data examined across different hierarchical levels starting from level 2 (Phylum) to level 6 (Genus). For LEfSe, differentially abundant features were determined using a non-parametric factorial Kruskal-Wallis (KW) rank-sum test to categorize statistical differences seen between two or more groups. After identifying statistical differences between groups, additional tests were done using the Wilcoxon rank-sum test and Linear Discriminant Analysis in order to normalize samples and assess if the differences were consistent with biological relevancies [19]. Using LEfSe, comparisons were made between two groups at any one time (e.g., Day 1 vs. Day 60, Farm A vs, Farm B, etc.). For STAMP, OTUs were assigned to different subsystems or biological pathways. Statistical tests were run on the samples at various hierarchical levels using effect sizes and confidence intervals to assess the biological importance of bacterial communities’ interactions with foals. A taxonomic abundance data file was submitted into STAMP along with a metadata file containing information such as foal/mare identity and time point. Default parameters were used with the addition of a Bonferroni multiple test correction. Data outputs from STAMP were input into R-studio; statistical outputs in a multiple group statistics table created by STAMP were inputted to create dendritic heatmaps [20]. Mean relative frequencies of each OTU in each fecal sample allowed for visualization of relatedness amongst the samples along with the abundance of specific bacteria in different groups.

Results and discussion

Diversity analysis

From the 163 samples, 11,793,830 reads were generated. In QIIME, sequence reads were segregated by sample, which represents a horse (mare or foal) at a time point after birth. For 3,102,798 reads, OTUs were picked based on sequence similarity using the Greengenes reference library; after filtering for a minimum of two counts per OTU, 2,447,151 reads had OTUs picked for a mean of 15,103 counts per sample (Table 1). A rarefaction curve was generated to compare level of diversity and depth of coverage by time point (Fig 1). Numbers of diverse OTU assignments were similar amongst Day 60 foals, weaned foals, and mares. The least diversity was seen at Day 7, and coverage depth was smallest for Day 1 foal fecal samples. The lack of a plateau on the rarefaction curve graphing taxonomic diversity versus sequencing depth demonstrated that saturation of observed OTUs was not met; however, increasing the threshold for minimal reads picked for OTUs above 3590 reads would have reduced the number of horses represented. Differences in alpha diversity most often demonstrated greater alpha diversity by age (S3 Table).

[Figure omitted. See PDF.]

Table 1. Sequencing Results.

https://doi.org/10.1371/journal.pone.0216211.t001

[Figure omitted. See PDF.]

Fig 1. Rarefaction curve compares diversity and depth coverage amongst samples segregated by time points.

As a general trend, the greatest depth of coverage by sample was seen for foals at Day 60 and weaning as well as in adult mares. This curve was used to determine the number of sequence reads to use as a threshold to capture more fecal samples in the study.

https://doi.org/10.1371/journal.pone.0216211.g001

Microbial populations from all fecal samples sequenced were subjected beta diversity analysis. An unweighted Unifrac PCA is represented in Fig 2. The compositions of bacterial populations for the samples follow a pattern throughout the early life of the foal. Day 1 and 7 samples, respectively, cluster according to their time points, regardless of farm. There is a transition of samples up to and through Day 28. Then Day 60, weaning, and adult samples (mare) cluster together. These data indicate that by Day 60 the foal’s GI microbiota has been established to include the bacteria necessary for the digestion of the roughage typically found in the mature horse diet, much like what had been shown in a study of Quarter Horses in Canada [5]. A dendritic heatmap of phyla level microbial population data shows a significantly high abundance of Proteobacteria, mainly of the Acinetobacter genus, in Day 1 foals compared to the other age groups (Fig 3). Proteobacteria continued to be present in the Day 7 samples and lowered significantly by Day 28. From age groups Day 7 to day of weaning and mares, abundance of Firmicutes followed by Bacteroidetes was observed throughout the samples with a similarity between the age groups up to the family level. Other studies have also shown that Firmicutes are the most abundant phylum in horses [5, 21]. When examining samples for each farm, farm-dependent clustering dependent on farm location was not observed in the dendritic heatmap; however, farm-to-farm comparisons were further explored using LDA Effect Size (LEfSe) analysis.

[Figure omitted. See PDF.]

Fig 2. PCA demonstrates how globally foal microbiota changes toward adult microbiota.

A principal component analysis of bacterial populations found from sequence data shows that the microbiota transitions toward the adult GI microbiota. Foal samples at Days 1, 7, 28, 60, and weaning, respectively, cluster by day with Day 60 and weaning samples overlapping with the adult mare. Loose clusters represent periods of variability. Segregation of samples by farm can be found in S1 Fig.

https://doi.org/10.1371/journal.pone.0216211.g002

[Figure omitted. See PDF.]

Fig 3. A dendritic plot and heatmap of the phyla present within each sample’s sequencing data demonstrate how individual fecal samples cluster together.

From the sequence data, the most obvious finding in the heatmap is that Day 1 samples all have bacteria of the Phylum Proteobacteria in common, while samples outside of Day 1 all share bacteria from the Phylum Firmicutes. At the bottom of the heatmap is a color code of the samples in which the colors designated for each day are the same as Fig 2. In this heatmap, Day 1 foal samples cluster together; samples from other days cluster with each other.

https://doi.org/10.1371/journal.pone.0216211.g003

Microbial populations were then examined in taxa bar plots; we detected differences in microbial populations by age and farm (Fig 4, S4 Table). Pseudomonas and Acinetobacter of the class Proteobacteria, including from family Moraxellaceae, were highly present in Day 1 fecal samples while Enterobacteriaceae, Fusobacterium, Erysipelotrichaceae, Peptostreptococcaceae, and Bacteroides were highly present in Day 7 fecal samples. A newborn’s gut represents a positive redox potential allowing for the growth of benefitting facultative anaerobes [22]. Pseudomonas' capability to rapidly multiply and its opportunistic nutritional strategy allow it to flourish in a foal’s initial gut colonization [23]. Along with Pseudomonas, other facultative anaerobes such as Fusobacterium and Peptostreptococcaceae can begin to predominate the intestinal microbiota [24]. Notably, previous studies have shown that numbers of Bacteroides rise sharply after the introduction of solid food [25]. At a week of age, we expect the foals to be progressively exposed to solid food through contact with their mother’s concentrate and forage corresponding to the higher abundance of these bacteria observed in day 7 foals [24]. Greater levels of Enterobacteriaceae at Day 7 are likely the result of exposure of the foal to the mare’s feces and copraphagic events affecting the foal’s microbiota [26, 27]. It should also be noted that two common microbial families in foals from Day 7 until weaning, and in mares, were Lachnospiraceae and Ruminococcaceae necessary for breakdown of complex carbohydrates like with grass diets [28]. Quercia et al. theorized bacteria from these families from the mares help with colonization of the foal GI [7].

[Figure omitted. See PDF.]

Fig 4. QIIME’s taxonomical plots can be used to compare and contrast across farm, age, and mare or foal status.

From the preliminary sequence data, certain trends can be seen. For example, in the first day of life fecal samples contain higher levels of bacteria from the Moraxellaceae family (*), commonly found in soil. By Day 7, there is a great deal of variation. Two common populations in foals from Day 7 until weaning, and in mares, include bacteria from families Lachnospiraceae (†) and Ruminococcaceae (‡) necessary for breakdown of complex carbohydrates like with grass diets.

https://doi.org/10.1371/journal.pone.0216211.g004

LEfSe analysis comparing populations.

LDA Effect Size (LEfSe) was used to classify differentially abundant microbiota features between subgroups of foals. After performing a KW sum-rank test to categorize statistically different levels of bacteria between groups, differentially abundant populations were found when comparing diarrheic and non-diarrheic samples, when comparing samples at different postnatal time points, and when comparing between the horses on the two major farms examined.

Ten of the 163 fecal samples (or 6.1%) in this study were diarrheic samples, representing incidence of diarrhea in eight foals (19.0%): one Day 1, four Day 7, two Day 28, two Day 60, and one at weaning. It should be noted that none of these foals had a failure of passive transfer of colostral immunoglobulins. Day 7 diarrheic samples were compared collectively to diarrheic samples at Days 1, 28, 60, and weaning (Fig 5A). Certain bacterial populations were greater in Day 7 relative to the other diarrheic samples. For example, for Day 7 diarrheic samples there was increased abundance of bacteria from the family Enterobacteriaceae including facultative anaerobes taking advantage of a newborn’s intestinal aerobic environment [29] and possibly representative of Salmonella species and Escherichia coli, as well as increased microbes from Alcaligenceae—though still <0.1%—which could represent the genus Sutterella which is linked to acute hemorrhagic diarrhea and inflammatory bowel disease [30]. Additionally, increased amounts of Bifidobacteriaceae bacteria were also present, likely representing the greater dependence of the Day 7 foal on milk relative to the older postnatal time points included in this comparison since microbes in this family are known to utilize lactose and milk oligosaccharides. Examination of some of the species seen in the other diarrheic samples could be indicative of dysbioses or just true differences by age such as Spirochaetes found in relatively higher numbers in healthy horses [31–33]. Compared to diarrheic Day 7 foal samples, Day 7 non-diarrheic samples were significantly enriched in order Actinomycetales especially the family Micrococcaceae (Fig 5B). Micrococcaceae and Actinomycetales have yet to be studied in detail within equine. In humans, members of the Actinomycetales order are associated with a healthy GI microbiota including the prevention of diarrhea, though Micrococcaceae are noted to be commensals [34, 35].

[Figure omitted. See PDF.]

Fig 5. LEfSe distinguishes difference in relative abundance of bacterial families in diarrheic samples at Day 7.

(A) Significant differences in microbial populations were examined for Day 7 diarrhea samples versus all other diarrhea samples. (B) Differences between microbial populations of Day 7 diarrhea samples versus non-diarrheic samples were also considered. Significant differences were found by Kruskal-Wallis (KW) rank-sum test, and reported in log 10 scale.

https://doi.org/10.1371/journal.pone.0216211.g005

Comparisons were also made with LEfSe between the two major farms at Days 7, 28, and 60. At Day 7, there were several microbial populations that were enriched in Farm A foals; these included Deltaproteobacteria, Desulfovibrionaceae, Veillonellaceae, Peptococcaceae, Pasteurellaceae, Odoribacteraceae, Erysipelotrichaceae, Thermomicrobia, Fusobacteriaceae, and Lachnospiraceae (Fig 6A). Interestingly, bacteria from Deltaproteobacteria, Desulfovibrionaceae, and Erysipelotrichaceae are found in greater abundance in high fat diets like with milk consumption [36, 37]; thus, as Farm B was exclusively Thoroughbred, perhaps there was more milk fat in the mare’s milk of the several breeds at Farm A. Moreover, microbes from Odoribacteraceae and Lachnospiraceae have been found to contribute to the production of short chain fatty acids, particularly butyrate which is used as an energy source for gut epithelial cells [38] and aids in combatting against GI disorders, such as Clostridial difficile infection [39]. Farm B Day 7 foals were significantly more abundant in Enterococcaceae and Fibrobacteraceae. Studies have discussed the initial colonization of the neonatal intestinal microbiome containing Enterococcaceae before rapidly decreasing as Firmicutes begin to become dominant [40]. In the family Fibrobacteraceae, there is common bacteria, Fibrobacter, known to digest fiber which may be the result of the foal’s exposure to its new local environment [41].

[Figure omitted. See PDF.]

Fig 6. LEfSe distinguishes which bacterial populations are relatively abundant between healthy samples at Farm A and Farm B at different time points.

Specifically, we examined (A) Day 7, (B) Day 28, (C) Day 60, and (D) mare samples at Day 1. Significant differences were found by Kruskal-Wallis (KW) rank-sum test, and reported in log 10 scale.

https://doi.org/10.1371/journal.pone.0216211.g006

In comparing the microbial populations of fecal samples at Day 28, there were proportional differences between classes of Firmicutes, Actinobacteria, and Bacteroidetes, and Proteobacteria (Fig 6B). The 4 phyla are commonly found to dominate the gut microbiome. Samples from Farm A had higher relative abundances of Actinobacteria, Bacteroidetes, and Proteobacteria than those of Farm B. The presence of certain subclasses of Actinobacteria is an indicator of a healthy gut given their association with immune response modulation to combat pathogens [42]. Bacteroidetes are associated with health benefits through the creation of short chain fatty acids via the digestion of complex sugars and proteins [43]. Microbes from Proteobacteria, Campylobacteraceae, and Erysipelotrichaceae were found in greater abundance in fecal samples from Farm A. Campylobacteraceae contain genera which are associated with equine GI disease [44]. Through examination of the Firmicutes phylum, Farm A contained higher levels of Bacillales and Erysipelotrichi, while Farm B contained higher levels of Clostridia including Clostridiales, and Lachnospiraceae. Bacillales and Erysipelotrichi have been associated to increase with a high-fat diet [45, 46]; perhaps at Day 28, Farm A’s foals were still relying more upon mare’s milk. Commensal Clostridia are essential for proper maintenance of the gut and participate in several processes including physiology, metabolism, and immune responses [47]. However, it has been proposed that high levels of Clostridiales contribute to inflammatory bowel syndrome symptoms through the increased concentrations of intestinal butyrate being produced [48].

For fecal samples from 60-day-old foals, Farm A contained a greater abundance of classes Corynebacteriales (Dietziaceae and Corynebacteriaceae) and Micrococcales (Micrococcaceae) from the phylum Actinobacteria and Bacillales (Staphylococcaceae) from the phylum Firmicutes (Fig 6C). Species of the family Dietziaceae have been suggested to have a novel fatty acid biosynthesis system while high levels of Corynebacteriaceae have been associated with breast milk bacterial communities [49, 50]. Staphylococcaceae species are capable of causing infections in multiple animals including equine, bovine, and swine, while species in the order Micrococcales are associated with soil-born decaying plants [34]. Farm B represented increased levels of Euryarchaeota and Proteobacteria. Microbes from the Phylum Euryarcheota include methanogens and sulfide-reducers, suggesting that at sixty days of age horses from Farm B were more likely utilizing pasture than a forage-grain diet, relative to Farm A [51]. This is indeed the case as the horses are kept on pasture with hay forage offered for Farm B while horses in Farm A were fed hay with very limited pasture. In comparing the bacteria in the feces of foals from Farms A and B, differences in fecal microbiotas for the horses on each farm could be associated with differences in management. However, more thorough and better controlled studies will need to be done to better resolve such relationships.

Microbial populations of fecal samples from mares at the day of their foals’ birth were also compared on Farms A and B (Fig 6D). As mentioned above, observations from these differences could lead to further better-controlled studies to resolve relationships between these microbial populations and farm-to-farm difference. There were more microbes from the families Bacteroidaceae and Spirochaetaceae in the feces of mares from Farm A, while there were more microbes from the families Streptococcaceae, Succinivibrionaceae, Methanocorpusculaceae, and Christensenellaceae in the feces of mares from Farm B. Microbes from the families Bacteroidaceae and Spirochaetaceae contribute to the core composition of the bacteria in the equine large intestine [32]. Others have noted that relative decreases in Bacteroidaceae have been observed in animals that are stressed [6, 52]. Decreased relative levels of Bacteroidaceae along increased relative abundance of bacteria from the family Succinivibrionaceae, which have been found to increase with stress in horses [53, 54], proffer another study for further consideration of diet, stress, and parturition, much like Mach et al. did comparing salivary cortisol levels during weaning [6]. Some bacterial species in the family Christensenellaceae have been found in leaner animals; as mares in Farm B were exclusively Thoroughbreds, the differences in abundance of bacteria from the family could be due to breed differences in prevalence of obesity and Equine Metabolic Syndrome [55, 56]. A comparison between the fecal samples of the mares also seems to offer insight into possible associations with differences in conditions at each farm.

In addition, specific comparisons were made between the fecal samples of foals at Day 1 and Day 60 which was a time point in this study in which the gut microbiome was closely similar to that of the fecal microbiota of adult horses (Fig 7). At Day 1, Cyanobacteria, Fusobacteria, and Proteobacteria were significantly more abundant. Day 60 samples were more abundant for Firmicutes, Bacteroidetes, Verrucomicrobia, Synergistetes, Spirochaetes, Fibrobacteres, Tenericutes, Planctomycetes, and Euryarchaeota. Firmicutes are involved in assisting with the digestion of insoluble fiber and hindgut fermentation [57]. Bacteroidetes are also known to assist in hindgut fermentation, but to a lesser degree than Firmicutes [57]. Verrucomicrobia also play a role in hindgut fermentation and contains bacteria in its phylum known to be associated with healthier metabolic status in humans [58, 59]. However, Verrucomicrobia have been noted to increase with chronic laminitis in horses [59]. Not much is currently known about the phylum Synergistetes, but one study has described a positive correlation between the amount of Synergistetes and the production of anti-inflammatory antibodies [60]. Fibrobacteres has been found in the large intestine of adult horses and may play a role in the digestion of cellulose and other fibers [32]. In mouse, the amount of Tenericutes in the gut microbiome decreases with an increase of fat in the diet which is not typical of an adult horse diet [61].

[Figure omitted. See PDF.]

Fig 7. LEfSe distinguishes which bacterial populations are relatively abundant between healthy samples at Day 1 vs. Day 60.

Significant differences in microbial populations were found when comparing fecal samples of foals in the first day of life which rely upon mare’s milk with fecal samples from the same foals at Day 60 that also consume grains, pasture, and forage. Significant differences were found by Kruskal-Wallis (KW) rank-sum test, and reported in log 10 scale.

https://doi.org/10.1371/journal.pone.0216211.g007

STAMP analysis.

Statistical analysis of taxonomic and functional profiles (STAMP) was used to characterize functional roles of microbial populations in host biology. At the phylum level after Bonferroni correction, 28 categories were considered to have differed significantly at some point in the transition of microbial populations from birth to weaning (Fig 8). Notably, as pertaining to the foal’s diet over time, categories like metabolic diseases (Fig 8A), biosynthesis of other secondary metabolites (Fig 8B), metabolism of amino acids (Fig 8C), carbohydrate metabolism (Fig 8D), amino acid metabolism (Fig 8D), nucleotide metabolism (Fig 8D), and lipid metabolism (Fig 8D) demonstrate proportional differences representative of the digestion of milk-rich nutrients in a neonatal diet transitioning toward a complex carbohydrate-rich grain and forage diet. Also interesting was the dramatic difference between microbial populations typically characterized in responding to exposure to foreign compounds like xenobiotics, as well as polyketides (e.g., naturally occurring insecticides, antibiotics, antifungals, etc.) and terpenoids found in plants, possibly in response to a foal’s first exposure to its environment after parturition. Some of these xenobiotic exposures are found when STAMP is used to consider differences at class level, e.g., styrene degradation, atrazine degradation, naphthalene degradation, etc. (S5 and S6 Tables).

[Figure omitted. See PDF.]

Fig 8. STAMP (STatistical Analysis of Metagenomic Profiles) was used to characterize functional roles of microbial populations in host biology by age.

After Bonferroni correction, twenty-eight categories were considered to have differed significantly at some point in the transition of microbial populations from birth to weaning. In the above panels, the categories are divided by proportion ranges of (A) 0.0%—0.25%, (B) 0.25% -1.5%, (C) 1.5%—8.0%, and (D) 5%—20%. Bars are represented as mean ± standard deviation.

https://doi.org/10.1371/journal.pone.0216211.g008

Considering storage and comparisons.

In this study, fecal samples were stored at -20°C. Many studies have suggested that storing samples in liquid nitrogen and/or -80°C could be considered the gold standard [62–64]. However, very few studies have tested -20°C conditions [65–67], though others employed these conditions for storing fecal samples and even soil samples for downstream microbial analyses [68–74]. For any study, post-collection biases can arise from many factors, including timing to transport samples, temperatures, and DNA preservation solutions [67, 75]. Nonetheless, two recent studies have determined that when all samples are stored in the same manner, the variability between samples prevailed over variability due to storage effects [65, 66]. Moreover, when specifically considering -80°C storage to seven other combinations of storage temperatures and DNA preservatives, the samples stored at -20°C and -80°C were most similar [66]. Clearly, for purposes of comparing studies, the -80°C storage conditions are most optimal. Thus, we acknowledge that our samples were stored at -20°C. Furthermore, when considering any studies characterizing microbiotas, it is essential that researchers examine the methods and metadata associated with those studies. Our comparisons with LEfSe considering individual features, as opposed to comparing multiple features at once, were more effective for understanding differences. It is likely that more complex and stratified studies would require many more horses and samples to better resolve relationships. Moreover, for specific management differences mentioned, when considering the effects they might have on a microbiota, it is necessary to have controlled studies with matching considerations for as many parameters as possible, even for conditions like matching diets, temporal range of events of parturition, and exposure to weather and precipitation.

Conclusions

By using next-generation sequencing of the V4 domain of fecal DNA from foals at Days 1, 7, 28, 60, and weaning, we found that bacterial population compositions followed a pattern throughout the early life of the foal in an age-dependent manner. Moreover, we were also able to recognize differences in microbial populations amongst diarrheic foals, and we were able to detect microbial population differences that suggest impacts from differences in management styles between the facilities caring for these foals, though more thorough studies are required. Our future efforts will include strategies to better discern the effects of less abundant bacterial populations that may be just as important to GI health. Knowledge of how associations between microbial population profiles and animal management strategies can be used to inform horse owners and facility managers on the effects of their decisions on GI health of their horse herds.

Supporting information

[Figure omitted. See PDF.]

S1 Fig. Principal Component Analysis Segregating Out farms.

When considering beta diversity (unweighted unifrac), an examination of where each farm segregates within the PCA shows how samples by farm are represented within the cluster.

https://doi.org/10.1371/journal.pone.0216211.s001

(PDF)

S1 Table. Mare and foal set information.

https://doi.org/10.1371/journal.pone.0216211.s002

(XLSX)

S2 Table. Farm management information.

https://doi.org/10.1371/journal.pone.0216211.s003

(XLSX)

S3 Table. QIIME alpha diversity comparison.

https://doi.org/10.1371/journal.pone.0216211.s004

(XLSX)

S4 Table. Percentage composition of microbial populations, by family (L5).

https://doi.org/10.1371/journal.pone.0216211.s005

(XLSX)

S5 Table. ST2—L3 CategorySTAMP (color gradient high to low, p<0.05).

https://doi.org/10.1371/journal.pone.0216211.s006

(XLSX)

S6 Table. ST2—L3 STAMP significant categories (p<0.05).

https://doi.org/10.1371/journal.pone.0216211.s007

(XLSX)

Acknowledgments

This study was supported by a grant from the UC Davis Center for Equine Health and funds provided by the UC Davis College of Agriculture & Environmental Sciences and the Agriculture Experiment Station at UC Davis. This is a project within the USDA NRSP8 Multi-State Research Group.

Citation: De La Torre U, Henderson JD, Furtado KL, Pedroja M, Elenamarie O, Mora A, et al. (2019) Utilizing the fecal microbiota to understand foal gut transitions from birth to weaning. PLoS ONE 14(4): e0216211. https://doi.org/10.1371/journal.pone.0216211

References

1. Becvarova I, Buechner-Maxwell V. Feeding the foal for immediate and long-term health. Equine Vet J. 2012;44:149–56.

2. O’Hara AM, Shanahan F. The gut flora as a forgotten organ. EMBO reports. 2006;7(7):688–93. pmid:16819463.

3. Hurley WL, Theil PK. Perspectives on immunoglobulins in colostrum and milk. Nutrients. 2011;3(4):442–74. pmid:22254105.

4. Bordin AI, Suchodolski JS, Markel ME, Weaver KB, Steiner JM, Dowd SE, et al. Effects of administration of live or inactivated virulent Rhodococccus equi and age on the fecal microbiome of neonatal foals. PloS one. 2013;8(6):e66640. Epub 2013/06/21. pmid:23785508.

5. Costa MC, Stampfli HR, Allen-Vercoe E, Weese JS. Development of the faecal microbiota in foals. Equine veterinary journal. 2015. pmid:26518456.

6. Mach N, Foury A, Kittelmann S, Reigner F, Moroldo M, Ballester M, et al. The Effects of Weaning Methods on Gut Microbiota Composition and Horse Physiology. Front Physiol. 2017;8:535. Epub 2017/08/10. pmid:28790932.

7. Quercia S, Freccero F, Castagnetti C, Soverini M, Turroni S, Biagi E, et al. Early colonisation and temporal dynamics of the gut microbial ecosystem in Standardbred foals. Equine veterinary journal. 2019;51(2):231–7. Epub 2018/06/23. pmid:29931762.

8. Faubladier C, Sadet-Bourgeteau S, Philippeau C, Jacotot E, Julliand V. Molecular monitoring of the bacterial community structure in foal feces pre- and post-weaning. Anaerobe. 2014;25:61–6. pmid:24315809.

9. Kuhl J, Winterhoff N, Wulf M, Schweigert FJ, Schwendenwein I, Bruckmaier RM, et al. Changes in faecal bacteria and metabolic parameters in foals during the first six weeks of life. Vet Microbiol. 2011;151(3–4):321–8. pmid:21511405

10. Blunden AS, Smith KC, Neal HN, Dugdale DJ. Enteritis associated with infection by Campylobacter jejuni in two Thoroughbred foals. Equine Vet Educ. 2006;18(1):8–10.

11. Pusterla N, Byrne BA, Hodzic E, Mapes S, Jang SS, Magdesian KG. Use of quantitative real-time PCR for the detection of Salmonella spp in fecal samples from horses at a veterinary teaching hospital. Vet J. 2010;186(2):252–5.

12. Slovis NM, Elam J, Estrada M, Leutenegger CM. Infectious agents associated with diarrhoea in neonatal foals in central Kentucky: A comprehensive molecular study. Equine Vet J. 2014;46(3):311–6. pmid:23773143

13. Anderson K. Nutrition Management of Pregnant and Lactating Mares. In: Extension Division IoAaNRatUoN-L, editor. Lincoln, NE: The Board of Regents of the University of Nebraska; 2011.

14. Horses" CoNRo. Nutrient Requirements of Horses, Sixth Revised Edition, 2007: National Academy of Sciences; 2007. 6th:[https://nrc88.nas.edu/nrh/.

15. Mon KK, Saelao P, Halstead MM, Chanthavixay G, Chang HC, Garas L, et al. Salmonella enterica Serovars Enteritidis Infection Alters the Indigenous Microbiota Diversity in Young Layer Chicks. Frontiers in veterinary science. 2015;2:61. pmid:26664988.

16. Bokulich NA, Subramanian S, Faith JJ, Gevers D, Gordon JI, Knight R, et al. Quality-filtering vastly improves diversity estimates from Illumina amplicon sequencing. Nature methods. 2013;10(1):57–9. pmid:23202435.

17. Caporaso JG, Kuczynski J, Stombaugh J, Bittinger K, Bushman FD, Costello EK, et al. QIIME allows analysis of high-throughput community sequencing data. Nat Methods. 2010;7(5):335–6. pmid:20383131

18. Parks DH, Tyson GW, Hugenholtz P, Beiko RG. STAMP: statistical analysis of taxonomic and functional profiles. Bioinformatics. 2014;30(21):3123–4. pmid:25061070.

19. Segata N, Izard J, Waldron L, Gevers D, Miropolsky L, Garrett WS, et al. Metagenomic biomarker discovery and explanation. Genome biology. 2011;12(6):R60. pmid:21702898.

20. Team R. RStudio: Integrated Development Environment for R. 1.0.136 ed2016.

21. Ericsson AC, Johnson PJ, Lopes MA, Perry SC, Lanter HR. A Microbiological Map of the Healthy Equine Gastrointestinal Tract. PloS one. 2016;11(11):e0166523. Epub 2016/11/16. pmid:27846295.

22. Adlerberth I, editor Establishment of a Normal Intestinal Microfiora in the Newborn Infant. Nestlé Nutrition Workshop Series; 1999: ROWEN PRESS.

23. Palleroni NJ. Practical Handbook The Genus Pseudomonas. In: Goldman E, Green LH, editors. Practical Handbook of Microbiology. 2nd ed. Boca Raton, FL:: CRC Press; 2008. p. 231–41.

24. Bezirtzoglou E. The intestinal microflora during the first weeks of life. Anaerobe. 1997;3(2–3):173–7. pmid:16887585

25. Rodriguez JM, Murphy K, Stanton C, Ross RP, Kober OI, Juge N, et al. The composition of the gut microbiota throughout life, with an emphasis on early life. Microb Ecol Health Dis. 2015;26:26050. Epub 2015/02/06. pmid:25651996.

26. Tanaka M, Nakayama J. Development of the gut microbiota in infancy and its impact on health in later life. Allergol Int. 2017;66(4):515–22. Epub 2017/08/23. pmid:28826938.

27. Crowell-Davis SL, Houpt KA. Coprophagy by foals: effect of age and possible functions. Equine veterinary journal. 1985;17(1):17–9. Epub 1985/01/01. pmid:4038939

28. Flint HJ, Scott KP, Duncan SH, Louis P, Forano E. Microbial degradation of complex carbohydrates in the gut. Gut microbes. 2012;3(4):289–306. pmid:22572875.

29. Arrieta M-C, Stiemsma LT, Amenyogbe N, Brown EM, Finlay B. The intestinal microbiome in early life: health and disease. Frontiers in immunology. 2014;5.

30. Lopetuso LR, Petito V, Graziani C, Schiavoni E, Sterbini FP, Poscia A, et al. Gut Microbiota in Health, Diverticular Disease, Irritable Bowel Syndrome, and Inflammatory Bowel Diseases: Time for Microbial Marker of Gastrointestinal Disorders. Digestive Diseases. 2017.

31. Costa MC, Weese JS. The equine intestinal microbiome. Animal Health Research Reviews. 2012;13(1):121–8. pmid:22626511

32. Dougal K, de la Fuente G, Harris PA, Girdwood SE, Pinloche E, Newbold CJ. Identification of a Core Bacterial Community within the Large Intestine of the Horse. PLoS ONE. 2013;8(10):e77660. pmid:24204908

33. Venable E, Bland S, McPherson J, Francis J. Role of the gut microbiota in equine health and disease. Animal Frontiers. 2016;6(3):43–9.

34. Hetem DJ, Rooijakkers SHM, Ekkelenkamp MB. 176—Staphylococci and Micrococci A2—Cohen, Jonathan. In: Powderly WG, Opal SM, editors. Infectious Diseases (Fourth Edition): Elsevier; 2017. p. 1509–22.e2.

35. Schell MA, Karmirantzou M, Snel B, Vilanova D, Berger B, Pessi G, et al. The genome sequence of Bifidobacterium longum reflects its adaptation to the human gastrointestinal tract. Proceedings of the National Academy of Sciences. 2002;99(22):14422–7.

36. Hildebrandt MA, Hoffmann C, Sherrill-Mix SA, Keilbaugh SA, Hamady M, Chen YY, et al. High-fat diet determines the composition of the murine gut microbiome independently of obesity. Gastroenterology. 2009;137(5):1716–24 e1–2. Epub 2009/08/27. pmid:19706296.

37. Koleva PT, Bridgman SL, Kozyrskyj AL. The infant gut microbiome: evidence for obesity risk and dietary intervention. Nutrients. 2015;7(4):2237–60. pmid:25835047

38. Gomez-Arango LF, Barrett HL, McIntyre HD, Callaway LK, Morrison M, Nitert MD. Increased Systolic and Diastolic Blood Pressure Is Associated With Altered Gut Microbiota Composition and Butyrate Production in Early PregnancyNovelty and Significance. Hypertension. 2016;68(4):974–81.

39. Schubert AM, Rogers MA, Ring C, Mogle J, Petrosino JP, Young VB, et al. Microbiome data distinguish patients with Clostridium difficile infection and non-C. difficile-associated diarrhea from healthy controls. MBio. 2014;5(3):e01021–14. pmid:24803517

40. Claus SP, Ellero SL, Berger B, Krause L, Bruttin A, Molina J, et al. Colonization-induced host-gut microbial metabolic interaction. MBio. 2011;2(2):e00271–10. Epub 2011/03/03. pmid:21363910.

41. Flint HJ, Bayer EA, Rincon MT, Lamed R, White BA. Polysaccharide utilization by gut bacteria: potential for new insights from genomic analysis. Nat Rev Microbiol. 2008;6(2):121–31. Epub 2008/01/09. pmid:18180751.

42. Ventura M, Canchaya C, Tauch A, Chandra G, Fitzgerald GF, Chater KF, et al. Genomics of Actinobacteria: tracing the evolutionary history of an ancient phylum. Microbiol Mol Biol Rev. 2007;71(3):495–548. Epub 2007/09/07. pmid:17804669.

43. Larsbrink J, Rogers TE, Hemsworth GR, McKee LS, Tauzin AS, Spadiut O, et al. A discrete genetic locus confers xyloglucan metabolism in select human gut Bacteroidetes. Nature. 2014;506(7489):498–502. Epub 2014/01/28. pmid:24463512.

44. Lastovica AJ, On SLW, Zhang L. The Family Campylobacteraceae. In: Rosenberg E, DeLong EF, Lory S, Stackebrandt E, Thompson F, editors. The Prokaryotes: Deltaproteobacteria and Epsilonproteobacteria. Berlin, Heidelberg: Springer Berlin Heidelberg; 2014. p. 307–35.

45. Antony K, Ma J, Prince A, Benjamin R, Cook C, Chu D, et al. 342: Poor maternal nutrition is associated with alterations in the maternal and neonatal microbiome communities. American Journal of Obstetrics & Gynecology. 212(1):S181.

46. Turnbaugh PJ, Ridaura VK, Faith JJ, Rey FE, Knight R, Gordon JI. The Effect of Diet on the Human Gut Microbiome: A Metagenomic Analysis in Humanized Gnotobiotic Mice. Science translational medicine. 2009;1(6):6ra14–6ra. pmid:20368178

47. Lopetuso LR, Scaldaferri F, Petito V, Gasbarrini A. Commensal Clostridia: leading players in the maintenance of gut homeostasis. Gut Pathogens. 2013;5:23-. pmid:23941657

48. Ferrario C, Taverniti V, Milani C, Fiore W, Laureati M, De Noni I, et al. Modulation of fecal Clostridiales bacteria and butyrate by probiotic intervention with Lactobacillus paracasei DG varies among healthy adults. The Journal of nutrition. 2014;144(11):1787–96. pmid:25332478

49. Nishiuchi Y, Baba T, Yano I. Mycolic acids from Rhodococcus, Gordonia, and Dietzia. Journal of microbiological methods. 2000;40(1):1–9. pmid:10739337

50. Bezirtzoglou E, Stavropoulou E. Immunology and probiotic impact of the newborn and young children intestinal microflora. Anaerobe. 2011;17(6):369–74. pmid:21515397

51. St-Pierre B, Cersosimo LM, Ishaq SL, Wright A-DG. Toward the identification of methanogenic archaeal groups as targets of methane mitigation in livestock animalsr. Frontiers in microbiology. 2015;6:776. pmid:26284054

52. Karl JP, Margolis LM, Madslien EH, Murphy NE, Castellani JW, Gundersen Y, et al. Changes in intestinal microbiota composition and metabolism coincide with increased intestinal permeability in young adults under prolonged physiological stress. Am J Physiol Gastrointest Liver Physiol. 2017;312(6):G559–G71. Epub 2017/03/25. pmid:28336545.

53. Destrez A, Grimm P, Julliand V. Dietary-induced modulation of the hindgut microbiota is related to behavioral responses during stressful events in horses. Physiol Behav. 2019;202:94–100. Epub 2019/02/07. pmid:30726719.

54. Warzecha CM, Coverdale JA, Janecka JE, Leatherwood JL, Pinchak WE, Wickersham TA, et al. Influence of short-term dietary starch inclusion on the equine cecal microbiome. Journal of animal science. 2017;95(11):5077–90. Epub 2018/01/03. pmid:29293739.

55. Biddle AS, Tomb JF, Fan Z. Microbiome and Blood Analyte Differences Point to Community and Metabolic Signatures in Lean and Obese Horses. Frontiers in veterinary science. 2018;5:225. Epub 2018/10/09. pmid:30294603.

56. Pleasant RS, Suagee JK, Thatcher CD, Elvinger F, Geor RJ. Adiposity, plasma insulin, leptin, lipids, and oxidative stress in mature light breed horses. Journal of veterinary internal medicine / American College of Veterinary Internal Medicine. 2013;27(3):576–82. Epub 2013/03/23. pmid:23517373.

57. Costa MC, Arroyo LG, Allen-Vercoe E, Stämpfli HR, Kim PT, Sturgeon A, et al. Comparison of the Fecal Microbiota of Healthy Horses and Horses with Colitis by High Throughput Sequencing of the V3-V5 Region of the 16S rRNA Gene. PLoS ONE. 2012;7(7):e41484. pmid:22859989

58. Dao MC, Everard A, Clément K, Cani PD. Losing weight for a better health: Role for the gut microbiota. Clinical Nutrition Experimental. 2016;6:39–58. http://dx.doi.org/10.1016/j.yclnex.2015.12.001.

59. Steelman SM, Chowdhary BP, Dowd S, Suchodolski J, Janečka JE. Pyrosequencing of 16S rRNA genes in fecal samples reveals high diversity of hindgut microflora in horses and potential links to chronic laminitis. BMC Veterinary Research. 2012;8:231-. pmid:23186268

60. López P, De Paz B, Rodríguez-Carrio J, Hevia A, Sánchez B, Margolles A, et al. Th17 responses and natural IgM antibodies are related to gut microbiota composition in systemic lupus erythematosus patients. Scientific reports. 2016;6:srep24072.

61. Everard A, Lazarevic V, Gaïa N, Johansson M, Ståhlman M, Backhed F, et al. Microbiome of prebiotic-treated mice reveals novel targets involved in host response during obesity. The ISME journal. 2014;8(10):2116. pmid:24694712

62. Dominianni C, Wu J, Hayes RB, Ahn J. Comparison of methods for fecal microbiome biospecimen collection. BMC Microbiol. 2014;14:103. Epub 2014/04/25. pmid:24758293.

63. Choo JM, Leong LE, Rogers GB. Sample storage conditions significantly influence faecal microbiome profiles. Sci Rep. 2015;5:16350. Epub 2015/11/18. pmid:26572876.

64. Roesch LF, Casella G, Simell O, Krischer J, Wasserfall CH, Schatz D, et al. Influence of fecal sample storage on bacterial community diversity. Open Microbiol J. 2009;3:40–6. Epub 2009/05/15. pmid:19440250.

65. Blekhman R, Tang K, Archie EA, Barreiro LB, Johnson ZP, Wilson ME, et al. Common methods for fecal sample storage in field studies yield consistent signatures of individual identity in microbiome sequencing data. Sci Rep. 2016;6:31519. Epub 2016/08/17. pmid:27528013.

66. Bundgaard-Nielsen C, Hagstrom S, Sorensen S. Interpersonal Variations in Gut Microbiota Profiles Supersedes the Effects of Differing Fecal Storage Conditions. Sci Rep. 2018;8(1):17367. Epub 2018/11/28. pmid:30478355.

67. Jenkins SV, Vang KB, Gies A, Griffin RJ, Jun SR, Nookaew I, et al. Sample storage conditions induce post-collection biases in microbiome profiles. BMC Microbiol. 2018;18(1):227. Epub 2018/12/29. pmid:30591021.

68. Arboleya S, Sanchez B, Milani C, Duranti S, Solis G, Fernandez N, et al. Intestinal microbiota development in preterm neonates and effect of perinatal antibiotics. J Pediatr. 2015;166(3):538–44. Epub 2014/12/03. pmid:25444008.

69. Diaz M, Guadamuro L, Espinosa-Martos I, Mancabelli L, Jimenez S, Molinos-Norniella C, et al. Microbiota and Derived Parameters in Fecal Samples of Infants with Non-IgE Cow’s Milk Protein Allergy under a Restricted Diet. Nutrients. 2018;10(10). Epub 2018/10/14. pmid:30314304.

70. Garas LC, Cooper CA, Dawson MW, Wang JL, Murray JD, Maga EA. Young Pigs Consuming Lysozyme Transgenic Goat Milk Are Protected from Clinical Symptoms of Enterotoxigenic Escherichia coli Infection. The Journal of nutrition. 2017;147(11):2050–9. Epub 2017/09/29. pmid:28954839.

71. Hansen TA, Joshi T, Larsen AR, Andersen PS, Harms K, Mollerup S, et al. Vancomycin gene selection in the microbiome of urban Rattus norvegicus from hospital environment. Evol Med Public Health. 2016;2016(1):219–26. Epub 2016/07/15. pmid:27412864.

72. Mariat D, Firmesse O, Levenez F, Guimaraes V, Sokol H, Dore J, et al. The Firmicutes/Bacteroidetes ratio of the human microbiota changes with age. BMC Microbiol. 2009;9:123. Epub 2009/06/11. pmid:19508720.

73. Penders J, Thijs C, van den Brandt PA, Kummeling I, Snijders B, Stelma F, et al. Gut microbiota composition and development of atopic manifestations in infancy: the KOALA Birth Cohort Study. Gut. 2007;56(5):661–7. Epub 2006/10/19. pmid:17047098.

74. Rubin BE, Gibbons SM, Kennedy S, Hampton-Marcell J, Owens S, Gilbert JA. Investigating the impact of storage conditions on microbial community composition in soil samples. PloS one. 2013;8(7):e70460. Epub 2013/08/13. pmid:23936206.

75. Horng KR, Ganz HH, Eisen JA, Marks SL. Effects of preservation method on canine (Canis lupus familiaris) fecal microbiota. PeerJ. 2018;6:e4827. Epub 2018/05/31. pmid:29844978.