Skip to main content
NCBI home page
Society for Reproduction and Fertility Open Access logoLink to Society for Reproduction and Fertility Open Access
. 2025 Jun 18;170(1):e250044. doi: 10.1530/REP-25-0044

Divergence in the sow vaginal microbiota is associated with fertility

Lauren Fletcher 1, Xiaoshu Zhan 1,2, Yashu Song 1, Julang Li 1,✉
PMCID: PMC12177788  PMID: 40470844

Abstract

In brief

Vaginal microbiota composition influences female fertility, however it has not been studied for measuring fertility level in female pigs. This study reveals significant vaginal microbiota composition differences between high reproductive performance and infertile sows, and demonstrates that the vaginal microbiota has promise for improving female pig selection using machine learning modeling.

Abstract

There is a need for reliable and effective biomarkers of female fertility and reproductive potential in the pork industry, as current selection protocols are not keeping up with the rate of improvement for other production-related traits. This study aimed to investigate the vaginal microbiota composition between sows of differing fertility status and identify candidate vaginal microbiota biomarkers of sow fertility. The vaginal microbiota of high reproductive performance sows (HRP, n = 52) with number of piglets born alive ≥13 and infertile sows (INF, n = 23), that remained nonpregnant after two consecutive rounds of artificial insemination, were investigated. Sequencing results revealed significantly different (P < 0.05) beta diversity at the genus level between HRP and INF vaginal microbiota communities. Accordingly, the composition of the vaginal microbiota diverged between HRP and INF sows, with INF sows having increased (P < 0.05) relative abundance of Lachnospiraceae XPB1014 group and HRP sows having increased (P < 0.05) relative abundance of Aerococcus and Staphylococcus at the genus level. Forty-two genera were selected as candidate biomarkers of sow fertility via partial least squares discriminant analysis (PLS-DA) and recursive feature elimination. The support-vector machine model classified sow fertility with 93.3% accuracy, supporting potential industry application to improve upon current methods for selection and recruitment in the breeding herd. Future investigations should validate the candidate vaginal microbiota biomarkers in a large, independent population of sows and gilts to evaluate their application for predicting future reproductive performance and assess their true industry applicability.

Keywords: fertility, vaginal microbiota, breeding herd, swine, machine learning

Introduction

The selection of female pigs for fertility and reproductive success is essential to promote sow lifetime productivity and retention in the breeding herd (Patterson & Foxcroft 2019). Genetics (Lillehammer et al. 2011, Knol et al. 2016) and metabolomics (Fletcher et al. 2022) are among the factors that can influence or reflect a female pig’s fertility and reproductive potential and have been explored for their application to commercial breeding herd selection. However, the composition of the pig vaginal microbiota has not been investigated in the context of female pig fertility status or for its capacity to aid breeding herd selection protocols through its association with female pig fertility or reproductive potential level. Given that the vaginal microbiota resides in the female reproductive tract, it likely plays a role in female pig conception and pregnancy success. In addition, due to its accessibility, it may be a convenient and valuable source of candidate biomarkers of fertility status.

The vaginal microbiome is a dynamic and interactive system with a mutualistic goal of host-microbiome homeostasis (Chen et al. 2021). Microbes that colonize the vagina associate with the vaginal epithelium for resources and secrete antimicrobial factors in return, such as bacteriocin, hydrogen peroxide or lactic acid, that help to maintain the health of the reproductive tract and to prevent the growth of pathogenic bacteria through competitive exclusion or killing (Kaewsrichan et al. 2006, Ghartey et al. 2014). The composition of the microbiota can be influenced by internal factors, including genetics, age, stage in the menstrual cycle and pregnancy, and external factors, such as diet, antibiotics, hygiene and microenvironment (Ravel et al. 2011). These factors can alter vaginal microbiome homeostasis, resulting in the loss of beneficial microbes, overgrowth of possibly harmful microbes and change in diversity. Accordingly, a wide range of negative reproductive states have been linked to shifts in human vaginal microbiota composition, including decreased IVF conception rates, increased risk of incidence of sexually transmitted infections, implantation failure, pregnancy loss, preterm birth and infertility (Fettweis et al. 2019, Haahr et al. 2019, Juliana et al. 2020, Grewal et al. 2022, Patel et al. 2022). Thus, the composition of the vaginal microbiota appears to be associated with fertility and reproductive health, suggesting that it may be an effective source of candidate biomarkers for fertility.

The literature on the female pig reproductive tract microbiome has been previously reviewed (Monteiro et al. 2022), highlighting investigations of urogenital infection (Wang et al. 2017, Zhang et al. 2021b), low and high reproductive performance after porcine reproductive and respiratory syndrome (PRRS) vaccination (Sanglard et al. 2020), and risk of pelvic organ prolapse in late gestation (Kiefer et al. 2021). Additional studies have investigated the vaginal microbiota between healthy sows and those with purulent discharge (Poor et al. 2022), sows with normal return-to-estrus and those without (Zhang et al. 2021a) and nonpregnant gilts and pregnant sows (Luque et al. 2021). However, vaginal microbiota composition relating to the fertility status of gilts or sows, specifically high reproductive performance and infertility, has not been investigated. Here, the composition of the vaginal microbiota of high reproductive performance (HRP) and infertile (INF) sows was investigated using 16S sequencing along with beta diversity, alpha diversity and relative abundance analysis. In addition, this study attempted to identify candidate vaginal microbiota biomarkers of sow fertility using a machine learning modeling approach.

Materials and methods

Animals and housing

All animal procedures were carried out in accordance with the Canadian Council on Animal Care (CCAC) guidelines under the Animal Utilization Protocol (AUP) #4825 approved by Animal Care Services (ACS) at the University of Guelph. Sows were either Yorkshire or Yorkshire-Landrace crosses that were individually housed in stalls or a group setting with access to water and restricted feed at Arkell Research Station (University of Guelph, Guelph, ON) or Alliance Genetics Canada’s affiliated operations. All animals included in the study were from 1 to 3 years of age, healthy, with no signs of illness or disease, and without abnormal vaginal discharge. HRP sows were between parity 1 and 6 with ≥13 piglets born alive on average throughout their reproductive lifespan immediately before sampling. Infertile sows (INF) were between parity 1 to 4 that remained nonpregnant after two rounds of back-to-back artificial insemination immediately before sampling (consecutive estrous cycles). All artificial inseminations were performed by experienced technicians using commercially sourced semen quality tested for sperm concentration, viability, motility, morphology and fertilization ability.

Sample collection

Vaginal microbiome samples were taken in triplicate from 77 sows (HRP: n = 53, INF: n = 24) fitting the above assessment criteria. HRP sows were sampled 4 to 5 days after weaning, an approximate time of commencement of estrus in sows (Soede et al. 2011). INF sows were sampled during their next expected estrous cycle after remaining nonpregnant for two heat cycles. Sampling occurred 21–23 days after their last detected estrus, or, 3 to 5 days post-injection if P.G. 600® was used for estrus synchronization. Estrus checks were performed following standard management practices using the standing response to back pressure test and visual changes in the vulva. All sows were indicated to be receptive for breeding by management when sampled. This was done to limit variability based on stage in the estrous cycle (Soede et al. 2011). Samples were also collected in the morning within the same 2-hour time window to limit variation based on circadian and diurnal variation. The outer vulva was washed with water and sprayed with 70% ethanol to ensure a pure vaginal microbiota sample collection. Swabs were inserted approximately 10–15 centimeters into the vagina, gently rotated for 20 s and immediately placed into PBS buffer on ice for transport, ensuring that the swab or vessel did not contact other surfaces. Upon arrival to the lab, the swabs were centrifuged at 9,000 ×g for 5 min to extract the bacterial biomass suspended in the swab. The recovered biomass was resuspended and added back into the PBS and bacteria solution, creating a homogenous sample suspension. A combined sample, an equal ratio of each replicate sample, was pooled and stored at −80°C until further analysis.

DNA extraction, 16S rDNA amplification and PacBio sequencing

Microbiome DNA extraction was performed using the Norgen Microbiome DNA Isolation Kit (Norgen Biotek, Canada) following manufacturer directions. Final extracted DNA was eluted with nuclease-free water. DNA was concentrated using the Thermo Scientific Savant™ SpeedVac™ DNA 130 Integrated Vacuum Concentrator System (Thermo Fisher Scientific, Canada) at ambient temperature to obtain a final volume of 20 μL.

The extracted microbiome DNA was amplified and prepared for 16S rRNA PacBio sequencing in a two-step PCR method. PCR1 was performed with primers Bact-27F (/5AmMC6/GCAGTCGAACATGTAGCTGACTCAGGTCACAGRGTTYGATYMTGGCTCAG) and Univ-1492R (/5AmMC6/TGGATCACTTGTGCAAGCATCACATCGTAGRGYTACCTTGTTACGACTT) (Frank et al. 2008), with universal PacBio adapters and relevant amino blocker modifications on the 5′ end to prevent ligation of the PCR product during SMRTbell library prep. The PCR1 primer sets were synthesized by Integrated DNA Technologies (IDT, USA). Each PCR1 reaction included 2 μL of the concentrated vaginal microbiome DNA, 2 μL 10× PCR buffer, 2 μL 50 mM MgCl2, 1.5 μL 2 mM dNTPs, 0.5 μL of 5 U/μL Taq Polymerase (Invitrogen™, USA), 1 μL 10 μM 27F primer, 1 μL 10 μM 1492R primer and 10 μL of molecular grade water for a final reaction volume of 20 μL. Samples were run on the CFX Connect Real-Time PCR Detection System Thermocycler (BioRad®, USA) under the following conditions: 94°C for 5 min, followed by 20 cycles of 94°C for 30 s, 65°C for 30 s and 72°C for 2 min and a final extension step at 72°C for 10 min. PCR1 products were stored at −20°C until further preparation. Test aliquots of random samples were visualized on 1% (w/v) agarose gel in SB buffer to ensure sufficient amplification before moving to the next step of the protocol.

PCR2, library preparation and sequencing were performed at the Centre for Biodiversity Genomics at the University of Guelph (Guelph, ON, Canada). PCR2 used PacBio unique molecular identifier (UMI) primers. The PCR2 cocktail for long PCR1 products per reaction was as follows: 13 μL 10% Trehalose, 2.8 μL 10× PCR Buffer, 1.3 μL 50 mM MgCl2, 1.3 μL 100% DMSO, 1.3 μL 0.2 mg/mL BSA, 0.5 μL of 10 μL of each UMI Primer (forward and reverse), 0.4 μL 10 mM dNTPs, 0.3 μL 5 U/μL Platinum Taq Polymerase (Invitrogen™) and 4.2 μL 10× diluted PCR1 DNA. The primers were added using Beckman Coulter’s Biomek FXP Liquid Handling Automation System. Samples were run on a thermocycler under the following conditions: 94°C for 2 min, 25 cycles at 94°C for 40 s, 64°C for 1 min, 72°C for 2 min and 72°C for 10 min. PCR2 products were quality checked on a 2% (w/v) agarose gel. Each reaction was pooled into the DNA library by adding 6 μL of each sample. The SMRTbell Prep Kit 3.0 (PN: 102-182-700, PacBio, USA) was used for library preparation with the procedure ‘Preparing multiplexed amplicon libraries using SMRTbell prep kit 3.0’ (PN: 102-359-000, PacBio) for Primer Barcoded samples, with minor adjustments to the protocol. Finally, the 16S rRNA amplicon libraries were sequenced on the Sequel IIe PacBio System with 15 h of movie time. A negative control sample was included throughout the protocol, undergoing all steps including sampling (swabs opened at location of pig sampling but unused), DNA extraction, PCR and sequencing.

Bioinformatics and statistical analysis

Amplicon libraries were demultiplexed using the standard SMRTlink demultiplexing procedure through Lima on HiFi reads with a minimum CCS Predicted Accuracy of 20 (phred score) and Barcode Score of 80. Demultiplexed reads were imported into R (version 4.2.3) (R Core Team 2023). The DADA2 package (version 1.26.0) (Callahan et al. 2016) was used to trim primers, filter by length (1,000–1,600 bp) and quality (Q > 3), dereplicate, denoise and generate amplicon sequence variants (ASVs). Samples with fewer than 1,500 filtered reads were removed. The DECIPHER R-package (version 2.26.0) (Wright 2016) was used to train a custom classifier to the species level using SILVA 138.2 (Mclaren & Callahan 2021, Quast et al. 2013) via LearnTaxa(). This classifier was used to assign taxonomy to the 16S sequencing data using the DECIPHER R-package function IdTaxa(). Unclassified ASVs and reads present in the negative control were removed. Classified reads were normalized by percent relative abundance at each taxonomic level. The phyloseq package (version 1.42.0) was used for relative abundance and genus-level alpha and beta diversity analyses (McMurdie & Holmes 2013) on the top 99% relatively abundant microbes. Relative abundance was examined at the phylum, genus and species levels to highlight broad and specific microbial composition patterns. Alpha and beta diversity analyses were performed at the genus level to optimize biological relevance with taxonomic resolution. This provides more meaningful insights than higher taxonomic levels, which may lack biological relevance due to result aggregation, and species level, which may lack taxonomic resolution accuracy. Beta diversity of the top 99% relatively abundant microbes was measured using Bray–Curtis dissimilarity (Sorenson 1948) and Jaccard distance (Jaccard 1908) via the ‘distance’ wrapper in the phyloseq package. Bray-Curtis is based on ASV abundance (Sorenson 1948) whereas Jaccard distance considers ASV presence or absence (Jaccard 1908). Resulting dissimilarity matrices were visualized with principal coordinates analysis (PCoA). Permutation Multivariate Analysis of Variance (PERMANOVA, 'adonis' function, 999 permutations) was used to quantify the multivariate community beta diversity between the experimental groups. The ‘betadisper’ function of the R vegan package (version 2.6–4) (Oksanen et al. 2022) confirmed homogeneity of group variances. Alpha diversity of the top 99% relatively abundant microbes was calculated using the abundance-based coverage estimator (ACE) metric (Chao & Lee 1992) for richness, Shannon’s Index (Shannon & Weaver 1949) for richness and evenness and Simpson’s Diversity Index (Simpson 1949) for richness and evenness by evaluating the probability that two randomly selected microbes belong to the same taxon. Measures were calculated using the ‘plot_richness’ function in phyloseq. Mann–Whitney U tests assessed univariate significance following homogeneity of variance (Levene test) and normality (Shapiro–Wilk test) tests. Benjamini–Hochberg adjusted P-values (False Discovery Rate; FDR) determined significant (P < 0.05) or trending toward significant (0.05 < P < 0.1) differences in vaginal microbe relative abundance.

Biomarker selection and performance analysis was completed using a modified version of a previously designed protocol (Fletcher et al. 2022), employing the Python library Scikit-learn (Pedregosa et al. 2011). Feature selection was performed using a combined approach with the supervised multivariate method Partial-Least Squares Discriminant Analysis (PLS-DA) and its associated variable importance of projection (VIP) metrics and recursive feature elimination (RFE) utilizing a decision tree classifier. Features with a VIP ≥1.5 in PLS-DA that were also selected in the RFE-CV() algorithm with the best accuracy across 5-fold cross-validation were selected as candidate biomarkers and used to train a support vector machine (SVM) classifier. A receiver operating characteristic (ROC) curve, depicting the performance of the trained SVM classifier, plotting false positive rate (100-specificity) against true positive rate (sensitivity) at various decision thresholds, was used. On the average ROC curve, the Youden’s J statistic (true positive rate - false positive rate) was calculated at each decision threshold to optimize true positive rate with false positive rate. The optimal decision threshold maximized Youden’s J statistic. Area under the curve (AUC) of the ROC curve and confusion matrix-based measures at the best decision threshold were used to evaluate the performance of the candidate biomarkers and trained SVM classifier.

Results

Out of the 77 sows sampled, 75 (HRP: n = 52, INF: n = 23) samples were included in the final analysis after quality control of sequencing results. A total of 686,906 high-quality reads were obtained from PacBio 16S sequencing of the vaginal microbiome samples, an average of approximately 9,159 reads per vaginal microbiome sample. The negative control sample had 1,001 filtered reads, with 99.5% of these reads being classified to the Pseudomonas genus, 0.24% being classified to the Bacillus genus and 0.24% being classified to the Phyllobacterium genus.

Relative abundance analysis

At the phylum level (Fig. 1A), Bacillota (41.7% in HRP, 41.2% in INF) was the most abundant in both HRP and INF microbiota. Pseudomonadota (26.4% in HRP, 19.9% in INF) was the second most abundant phylum in HRP microbiota, whereas Bacteroidota (16.7% in HRP, 24.0% in INF) was the second most abundant phylum in INF microbiota. Campylobacterota (13.2% in HRP, 13.1% in INF) was the third most abundant phylum in both HRP and INF microbiota. However, none of the differences in phylum-level relative abundance were significant between fertility groups (P > 0.1).

Figure 1.

Figure 1

Open in a new tab

Relative abundance of vaginal microbes in HRP and INF sows at the phylum (A), genus (B) and species (C) levels. Overall, shifts in abundance of the microbes composing the vaginal microbiota communities were apparent between HRP and INF sows. Microbes with a relative abundance of ≥1% are included; the rest (relative abundance <1%) are compiled together in ‘Other’.

At the genus level (Fig. 1B), the most abundant genera in HRP microbiota were Escherichia-Shigella (18.8%), Campylobacter (13.7%), Porphyromonas (11.2%), Clostridium (5.4%) and Veillonella (3.8%). In INF microbiota, Porphyromonas (18.8%) was the most abundant genus, followed by Campylobacter (13.8%), Escherichia-Shigella (10.7%), Streptococcus (5.7%) and Clostridium (4.1%). Interestingly, several genera were present at a relative abundance of 1% or greater in only one group: Aerococcus, Anaerococcus, Facklamia, Parvimonas, Pasteurella, Romboustia and Terrisporobacter in HRP and Actinobacillus, Comamonas, Prevotellaceae NK3B31 group, Rikenellaceae RC9 gut group and Ruminococcus in INF. In addition, Aerococcus and Staphylococcus were significantly (P < 0.05) increased in HRP vaginal microbiota compared to INF (Table 1). In contrast, Lachnospiraceae XPB1014 group was significantly increased (P < 0.05) and Anaeroplasma showed a trend toward a significant increase (0.05 < P < 0.1) in INF vaginal microbiota compared to HRP (Table 1).

Table 1.

Taxa with significantly (P < 0.05) and trending toward significantly (0.05 < P < 0.1) different relative abundance between HRP and INF sow vaginal microbiota communities. Bold P values indicate significantly different relative abundance between HRP and INF vaginal microbiota communities.

Taxonomic level/name Corrected P value Group it is increased Mean relative abundance %
HRP INF
Genus
 Lachnospiraceae XPB1014 group 0.005 INF 0.19 ± 0.51 1.01 ± 1.62
 Aerococcus 0.024 HRP 1.86 ± 2.83 0.18 ± 0.25
 Staphylococcus 0.026 HRP 1.02 ± 1.68 0.10 ± 0.15
 Anaeroplasma 0.084 INF 0.00 ± 0.00 0.07 ± 0.16
Species
 Campylobacter ureolyticus 0.009 HRP 13.39 ± 24.73 0.00 ± 0.00
 Staphylococcus simulans 0.009 HRP 1.67 ± 3.20 0.04 ± 0.16
 Corynebacterium urealyticum 0.014 INF 0.01 ± 0.08 0.28 ± 0.80
 Ruminococcus flavefaciens 0.014 INF 2.01 ± 8.14 8.32 ± 18.97
 Streptococcus alactolyticus 0.014 INF 0.65 ± 1.28 10.84 ± 18.05
 Streptococcus thoraltensis 0.017 HRP 1.19 ± 2.79 0.00 ± 0.00
 Streptococcus suis 0.018 HRP 5.95 ± 11.88 0.54 ± 1.39
 Facklamia tabacinasalis 0.026 HRP 7.16 ± 12.70 0.77 ± 1.75
 Peptoniphilus harei 0.026 HRP 0.55 ± 1.08 0.01 ± 0.04
Open in a new tab

At the species level (Fig. 1C), the most abundant species in HRP microbiota were Campylobacter ureolyticus (13.4%), Campylobacter hominis (10.2%), Porphyromonas somerae (9.7%), Campylobacter corcagiensis (8.4%) and Pasturella aerogenes (8.2%). In INF microbiota, Porphyromonas somerae (24.9%), Campylobacter corcagiensis (14.0%), Streptococcus alactolyticus (10.8%), Ruminococcus flavefaciens (8.3%) and Campylobacter hominis (6.2%) were the most prevalent. Interestingly, several species were present at a relative abundance of 1% or greater in only one group: Actinobacillus suis, Aerococcus suis, Bacteroides fragilis, Campylobacter lanienae, Campylobacter lanienae, Facklamia tabacinasalis, Staphylococcus simulans and Streptococcus suis in HRP and Streptococcus thoraltensis, Clostridium perfringens, Comamonas kerstersii, Corynebacterium xerosis, Finegoldia magna, Fusobacterium gastrosuis, Howardella ureilytica, Peptococcus simiae, Peptoniphilus grossensis, Streptococcus alactolyticus and Streptococcus dysgalactiae in INF. Moreover, Campylobacter ureolyticus, Staphylococcus simulans, Streptococcus thoraltensis, Streptococcus suis, Facklamia tabacinasalis and Peptoniphilus harei were all significantly increased (P < 0.05) in HRP vaginal microbiota. Corynebacterium urealyticum, Ruminococcus flavefaciens and Streptococcus alactolyticus were significantly increased (P < 0.05) in INF vaginal microbiota (Table 1).

Genus-level beta and alpha diversity analysis

The sow vaginal microbiota exhibits distinct diversity based on reproductive potential. Genus-level Bray-Curtis (PERMANOVA, F = 2.130, R2 = 0.028, P = 0.042) and Jaccard (PERMANOVA, F = 1.747, R2 = 0.0234, P = 0.041) beta diversity measures were significantly different between the vaginal microbiota of HRP and INF sows (Fig. 2A and B). When examining the top 15 taxa that contributed most to the PERMANOVA covariates (Fig. 2C), Escherichia-Shigella, Porphyromonas, Pasteurella and Streptococcus were the top contributing genera. In addition, the number of unique taxa observed and alpha diversity (ACE index, Shannon and Simpson) were examined at the genus level, with no significant differences (P > 0.10) based on reproductive potential (Fig. 3).

Figure 2.

Figure 2

Open in a new tab

Principal coordinate analysis (PCoA) of HRP and INF sow vaginal microbiota beta diversity at the genus level based on Bray-Curtis (A) and Jaccard (B) distance metrics. HRP and INF sows displayed significant (P < 0.05) beta diversity divergence at the genus level. The top 15 taxa contributing to the genus community differences were also explored (C).

Figure 3.

Figure 3

Open in a new tab

No significant alpha diversity differences were found between HRP and INF sow vaginal microbiota at the genus level. HRP and INF groups exhibited similar number of unique taxa observed and alpha diversity measures (ACE, Shannon and Simpson) at the genus level (P > 0.1).

Genus-level biomarker selection and model performance

PLS-DA distinguished between HRP and INF sows (Fig. 4) with moderate accuracy (73.3% ± 4.2), F1 score (75.4% ± 2.1), precision (80.5% ± 2.4) and recall (73.3% ± 4.2) across 5-fold cross-validation. The three PLS-DA components explained 37.8, 42.1 and 20.0% of the variance in genus-level vaginal microbiota features that contributed to distinguishing HRP and INF sows, respectively. The RFE model performed best when permitted to select 401 features, with 65.8% ± 15.4 accuracy. Combining these two complementary approaches using our feature selection protocol, 42 genera were selected to test as candidate vaginal microbiota biomarkers of fertility (Table 2). These candidate genus-level biomarkers performed well in the trained SVM classifier, achieving an overall AUC of 85.0% ± 9.0 across the five cross-validations in the ROC curve (Fig. 5). The best decision threshold that optimized true positive rate and false positive rate was 0.40. At this best decision threshold, the model achieved an overall accuracy of 93.3%, correctly classifying 94.2% of HRP sows and 91.3% of INF sows into their correct fertility groups.

Figure 4.

Figure 4

Open in a new tab

Genus-level feature selection using PLS-DA highlights distinct differences in HRP and INF vaginal microbiota composition. PLS-DA of genus-level taxonomic data displays differences in vaginal microbiota composition through a three-component PLS-DA, achieving moderate accuracy (73.3% ± 4.2), F1 Score (75.4% ± 2.1), precision (80.5% ± 2.4) and recall (73.3% ± 4.2) across 5-fold cross-validation. The percentage of explained variance for components 1, 2 and 3 was 37.82, 42.14 and 20.04%, respectively. This reflects the proportion of variance in the genus-level vaginal microbiota features captured to separate the HRP and INF groups.

Table 2.

List of 42 candidate genus-level vaginal microbiota biomarkers selected via PLS-DA and RFE analysis used for training the SVM algorithm to predict sow reproductive potential.

Candidate genus-level vaginal microbiota biomarkers
Lachnospiraceae XPB1014 group
Turicibacter
Proteiniclasticum
Corynebacterium
Terrisporobacter
Brumimicrobium
Lacticaseibacillus
Kocuria
Porphyromonas
Romboutsia
Salinicoccus
Sphingobacterium
Flaviflexus
Negativibacillus
Deinococcus
Staphylococcus
Weissella
Streptococcus
Anaeroplasma
dgA-11 gut group
Lactococcus
Facklamia
Enteractinococcus
Glaesserella
Treponema
Oscillospira
Kurthia
Amylolactobacillus
Aerosphaera
Tessaracoccus
Clostridium
Prevotella
Lapidilactobacillus
Aerococcus
Proteus
Quinella
Bifidobacterium
Cloacibacillus
Moraxella
Howardella
Yaniella
Ornithobacterium
Open in a new tab

Figure 5.

Figure 5

Open in a new tab

Performance of the SVM genus-level vaginal microbiota model across various decision thresholds, visualized by a ROC curve and corresponding AUC. The vaginal microbiota SVM model, trained on 42 candidate genus-level vaginal microbiota biomarkers, appeared diagnostic (AUC = 0.85 ± 0.09). The best decision threshold, determined by maximizing true positive rate and minimizing false positive rate, was 0.40, indicated by the red dot on the average ROC curve. The red line depicts a classifier with no predictive ability (AUC = 0.5); the blue line depicts the average predictive ability of the classifier over 5-fold stratified cross-validation.

Discussion

HRP and INF sows exhibit distinct vaginal microbiota composition, with INF communities shifting toward anaerobic and potentially harmful bacteria

The relative abundance of the top phyla observed in our pig vaginal microbiota samples appears to mirror past investigations of the pig vaginal microbiota. Typically, pig vaginal microbiota contain Pseudomonadota, Bacillota and Bacteroidota as the top three abundant phyla (Wang et al. 2017, Sanglard et al. 2020, Zhang et al. 2021a, Kiefer et al. 2021, Luque et al. 2021, Poor et al. 2022). When comparing the distribution of these phyla, HRP and INF sows had a similar relative abundance of Bacillota, but INF vaginal microbiota appeared to have decreased relative abundance of Pseudomonadota and increased Bacteroidota in comparison to HRP vaginal microbiota, although these shifts were not significant. Interestingly, decreased Pseudomonadota (Poor et al. 2022) in the vaginal microbiota of reproductively unhealthy pigs has been previously reported. In addition, in human cases of vaginosis, an increase in Bacteroidota has been noted (Wang et al. 2020). It appears that, at higher taxonomic levels, HRP and INF sow vaginal microbiota exhibit similar composition and distribution. Thus, the significant differences in composition observed may be driven by variations in the abundance of specific genus- or species-level features.

Typically, the vaginal microbiome is a relatively anaerobic environment, acquiring its oxygen, glucose and nutrients from the submucosal tissues (Linhares et al. 2011). Due to this environment, lactic acid is a major by-product of the anaerobic metabolism of many colonizing microbes using glycogen as an energy source (Linhares et al. 2011). However, during vaginal microbiome dysbiosis, the environment becomes increasingly anaerobic, with reduced lactic acid and glycogen availability (Mirmonsef et al. 2012, Tachedjian et al. 2017). This is associated with colonization by a mix of opportunistic anaerobic bacteria (Marrazzo 2011, Tachedjian et al. 2017) that produce mixed short-chain fatty acid (SCFA) profiles (Mirmonsef et al. 2011, Mirmonsef et al. 2012, Aldunate et al. 2015). These SCFAs are thought to contribute to vaginal microbiome dysbiosis (Amabebe & Anumba 2020) as they have pro-inflammatory effects in the vagina (Marrazzo 2011, Mirmonsef et al. 2011).

Consistent with this notion, the strict obligate anaerobic genera Anaeroplasma and Lachnospiraceae XPB1014 group appeared elevated in INF sow vaginal microbiota, with Anaeroplasma trending toward significance and Lachnospiraceae XPB1014 group exhibiting a significant increase. These genera could play a role in fostering a dysbiotic vaginal microbiome environment, as Anaeroplasma has been positively correlated with SCFA concentration (Wang et al. 2024) and Lachnospiraceae is known for SCFA production in the gut microbiome (Fusco et al. 2023), but it remains unknown if these roles transfer to the vaginal environment. In addition, the Lachnospiraceae genus has been found at higher abundance in low reproductive performance sows based on number of stillborn piglets (Sanglard et al. 2020). In HRP vaginal microbiota, Staphylococcus and Aerococcus were significantly increased in comparison to INF vaginal microbiota. These facultatively anaerobic genera have the ability to function in a healthy aerobic vaginal environment; however, information regarding their SCFA production in the vaginal microbiome is absent. Previous research has reported higher relative abundance of the Aerococcus genus in the vaginal microbiota of sows with healthy vaginal discharge compared to those with unhealthy purulent discharge (Poor et al. 2022). The Staphylococcus genus also appears to be a component of healthy pig vaginal flora and was found to be more abundant in sows of HRP based on number of stillborn and mummified piglets (Sanglard et al. 2020). It is possible that the INF vaginal microbiota may be shifting to include harmful bacteria that function in a strict anaerobic environment and produce SCFAs that play negative roles in the health of the vaginal environment.

Some striking shifts in relative abundance at the species level were observed between HRP and INF vaginal communities. A significantly higher abundance of Streptococcus alactolyticus and Ruminococcus flavefaciens in INF vaginal microbiota was observed at the species level. Streptococcus alactolyticus has been isolated from the vaginal microbiota of sows (Moreno et al. 2016, Renzhammer et al. 2020); however, details on its role in the vaginal microbiome and its impact on fertility remain unknown. In addition, an increase in the species Ruminococcus flavefaciens in the gut microbiome has been associated with sows that did not return to estrus (Zhang et al. 2021a); however, this relationship remains to be confirmed in the vaginal microbiota. A significantly higher abundance of Streptococcus suis, Facklamia tabacinasalis and Campylobacter ureolyticus was observed in HRP vaginal microbiota. Interestingly, Streptococcus suis is a very important and researched respiratory pathogen in the swine industry (Feng et al. 2014); however, it is known to be a component of a healthy pig vaginal microbiota (Murase et al. 2019), indicating that it may play a divergent and possibly beneficial role in the vaginal microbiome. Facklamia is a major lactic acid-producing genus which may enhance the competitive exclusion of harmful bacteria by maintaining a healthy pH and low concentrations of other inflammatory SCFAs (Björkroth & Koort 2011, Hoyles 2014). Although the Facklamia tabacinasalis species has been highlighted in the vaginal microbiota in a study focusing on pelvic organ prolapse in sows (Kiefer et al. 2021), its involvement in fertility and how it plays a role in the vaginal microbiome remain a question. Campylobacter ureolyticus is a species known to colonize the human vaginal ecosystem and enhance the growth of Lactobacillus species (Fernández-Edreira et al. 2023). Campylobacter ureolyticus was found to be increased in sows with a low risk of pelvic organ prolapse (Kiefer et al. 2021); however, research on this specific species of Campylobacter in the sow vaginal microbiota is limited.

Consistent with the statistically significant differences observed in relative abundance, beta diversity measures of HRP and INF sow vaginal microbiota were also statistically significant based on Bray-Curtis dissimilarity and Jaccard distance PERMANOVA analysis. In contrast, HRP and INF sow vaginal microbiota did not appear to have significantly different alpha diversity at the genus level. These results align with previous studies investigating the pig vaginal microbiota in the context of negative reproductive conditions, reporting significant composition differences (beta diversity) and limited or no differences in richness and evenness (alpha diversity) (Sanglard et al. 2020, Zhang et al. 2021a, Kiefer et al. 2021, Poor et al. 2022). Collectively, these findings support that the vaginal microbiota of HRP and INF sows diverge primarily in bacterial composition rather than overall taxa distribution.

Overall, the differences in vaginal microbiota composition between HRP and INF sows may influence vaginal microbiome homeostasis and contribute to reproductive condition or fertility status. However, whether the observed alterations in vaginal microbiota composition are causes or consequences of sow reproductive condition remains unclear. Furthermore, many of the highlighted genera and species are understudied in the context of the pig vaginal microbiome, leaving their specific roles to be determined. Further research is needed to clarify the roles of these microbes in female pig fertility, particularly to investigate whether they exert beneficial or harmful effects on the vaginal microbiome and in what capacity they have their effect.

SVM classifier accurately classifies sow fertility based on genus-level vaginal microbiota composition

The combined feature selection approach using PLS-DA and RFE selected 42 candidate biomarkers that enabled the SVM model to perform with strong classification performance. The trained SVM model demonstrated strong diagnostic performance for both fertility groups, achieving an area under the ROC curve of 85% ± 9.0 and overall accuracy of 93.3% at the best decision threshold. These results indicate that the model effectively discriminated between HRP and INF sows using the relative abundance data of the 42 genus-level vaginal microbiota biomarkers provided. This reinforces the importance of these selected genera in the pig vaginal microbiota and their association with sow fertility. Based on these results, it is possible that these genus-level vaginal microbes along with machine learning techniques could be applied to estimate female pig fertility in the breeding herd to make more informed management decisions. In addition, the vaginal microbiota could have predictive qualities for estimating future reproductive performance. This poses a valuable opportunity, as current breeding herd selection methods are resulting in a slower rate of genetic progress for vital litter and reproductive performance traits (Popovac et al. 2012, Lukac et al. 2016). Accordingly, breeding herd culling rates of 42.2–53.2% have been reported in North America (Ketchem et al. 2018) and this appears to stem from gilts in poor reproductive condition or sows with suboptimal first or second parity outcomes due to issues such as infertility, anestrus or poor reproductive performance (Hadaš et al. 2015, Kraeling & Webel 2015). High culling rates raise welfare concerns by increasing the number of gilts and sows needed to maintain normal production. In addition, this disturbs the flow of gilts and sows through the breeding herd, interferes with piglet production and results in an overall decrease of productivity and economic losses. An accurate vaginal microbiota model for accurately measuring fertility or reproductive potential could dramatically reduce culling related to suboptimal reproductive performance in the breeding herd, mitigating the economic and welfare impacts of current selection methods. However, future work should investigate these biomarkers in a larger population of gilts and sows to validate industry applicability and determine if the vaginal microbiota can be utilized to predict future reproductive performance before breeding herd recruitment.

Conclusion

The findings of this study suggest that sow vaginal microbiota composition is associated with fertility status, with INF sow microbiota exhibiting a potential dominance of anaerobic microbes with harmful tendencies that may outcompete the more aerobic and beneficial microbes found in HRP sow microbiota. Substantially more research is required to define the composition of a ‘typical’ healthy pig vaginal microbiota and to better understand how its bacterial components contribute to vaginal microbiome homeostasis. Several bacterial candidates that warrant further investigation for their roles in pig vaginal microbiome function and reproductive health have been highlighted. This study also demonstrates that the vaginal microbiota can be used to accurately classify sow fertility, with 42 candidate genus-level biomarkers highlighted. Accordingly, the composition of the vaginal microbiota could serve as a tool to inform breeding herd management decisions, including the selection of female pigs for the breeding herd, enhancing efficiency by identifying infertile or low-reproductive potential female pigs before several unsuccessful rounds of AI. However, for practical on-farm application and to confirm the true predictive value of these biomarkers, further validation in a larger independent population of sows and gilts, particularly in future reproductive performance, is essential.

Declaration of interest

The authors declare no conflicts of interest that could be perceived as prejudicing the impartiality of the research reported.

Funding

This work was supported by Food from Thought and the Canadian First Research Excellence Fund (grant numbers 499093 and 499199).

Author contribution statement

JL and LF conceived the study. JL secured funding, contributed to methodology design, assisted with data analysis and reviewed the manuscript. LF led methodology design, sample collection, sample and data analysis and manuscript preparation. XZ and YS assisted with sample collection and reviewed the manuscript.

Data availability

All relevant sequencing data included in this study is available in the Agri-environmental Research Data Repository at the University of Guelph: https://doi.org/10.5683/SP3/C6Y2LN.

References

  1. Aldunate M, Srbinovski D, Hearps AC, et al. 2015. Antimicrobial and immune modulatory effects of lactic acid and short chain fatty acids produced by vaginal microbiota associated with eubiosis and bacterial vaginosis. Front Physiol 6 164. ( 10.3389/fphys.2015.00164) [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Amabebe E & Anumba DOC. 2020. Female gut and genital tract microbiota-induced crosstalk and differential effects of short-chain fatty acids on immune sequelae. Front Immunol 11 2184. ( 10.3389/fimmu.2020.02184) [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Björkroth J & Koort J. 2011. Lactic acid bacteria | taxonomy and biodiversity. In Encyclopedia of Dairy Sciences, 2nd edn. pp 45–48. Ed Fuquay JW. San Diego, CA, USA: Academic Press. ( 10.1016/B978-0-12-374407-4.00255-7) [DOI] [Google Scholar]
  4. Callahan BJ, Mcmurdie PJ, Rosen MJ, et al. 2016. DADA2: high-resolution sample inference from Illumina amplicon data. Nat Methods 13 581–583. ( 10.1038/nmeth.3869) [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Chao A & Lee SM. 1992. Estimating the number of classes via sample coverage. J Am Stat Assoc 87 210–217. ( 10.2307/2290471) [DOI] [Google Scholar]
  6. Chen X, Lu Y, Chen T, et al. 2021. The female vaginal microbiome in health and bacterial vaginosis. Front Cell Infect Microbiol 11 631972. ( 10.3389/fcimb.2021.631972) [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Feng Y, Zhang H, Z ZW, et al. 2014. Streptococcus suis infection: an emerging/reemerging challenge of bacterial infectious diseases? Virulence 5 477–497. ( 10.4161/viru.28595) [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Fernández-Edreira D, Liñares-Blanco J, V-Del-Río P, et al. 2023. VIBES: a consensus subtyping of the vaginal microbiota reveals novel classification criteria. Comput Struct Biotechnol J 23 148–156. ( 10.1016/j.csbj.2023.11.050) [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Fettweis JM, Serrano MG, Brooks JP, et al. 2019. The vaginal microbiome and preterm birth. Nat Med 25 1012–1021. ( 10.1038/s41591-019-0450-2) [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Fletcher L, Akhtar N, Zhan X, et al. 2022. Identification of candidate salivary, urinary and serum metabolic biomarkers for high litter size potential in sows (Sus scrofa). Metabolites 12 1045. ( 10.3390/metabo12111045) [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Frank JA, Reich CI, Sharma S, et al. 2008. Critical evaluation of two primers commonly used for amplification of bacterial 16S rRNA genes. Appl Environ Microbiol 74 2461–2470. ( 10.1128/aem.02272-07) [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Fusco W, Lorenzo MB, Cintoni M, et al. 2023. Short-chain fatty-acid-producing bacteria: key components of the human gut microbiota. Nutrients 15 2211. ( 10.3390/nu15092211) [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Ghartey JP, Smith BC, Chen Z, et al. 2014. Lactobacillus crispatus dominant vaginal microbiome is associated with inhibitory activity of female genital tract secretions against Escherichia coli. PLoS One 9 e96659. ( 10.1371/journal.pone.0096659) [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Grewal K, Lee YS, Smith A, et al. 2022. Chromosomally normal miscarriage is associated with vaginal dysbiosis and local inflammation. BMC Med 20 38. ( 10.1186/s12916-021-02227-7) [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Haahr T, Zacho J, Bräuner M, et al. 2019. Reproductive outcome of patients undergoing in vitro fertilisation treatment and diagnosed with bacterial vaginosis or abnormal vaginal microbiota: a systematic PRISMA review and meta-analysis. BJOG 126 200–207. ( 10.1111/1471-0528.15178) [DOI] [PubMed] [Google Scholar]
  16. Hadaš Z, Schild M & Nevrkla P. 2015. Analysis of reasons for culling of sows in production herd. Res Pig Breed 9 1–5. [Google Scholar]
  17. Hoyles L 2014. The genus Facklamia. In Lactic Acid Bacteria: Biodiversity and Taxonomy, pp 91–98. Eds Holzapfel WH & Wood BJB. Hoboken, NJ, USA: Wiley. ( 10.1002/9781118655252.ch8) [DOI] [Google Scholar]
  18. Jaccard P 1908. Nouvellesrecherches sur la distribution florale. Bull Soc V Sci Nat 44 (163) 223–270. ( 10.5169/seals-268384) [DOI] [Google Scholar]
  19. Juliana NCA, Suiters MJM, Al-Nasiry S, et al. 2020. The association between vaginal microbiota dysbiosis, bacterial vaginosis, and aerobic vaginitis, and adverse pregnancy outcomes of women living in Sub-Saharan Africa: a systematic review. Front Public Health 8 567885. ( 10.3389/fpubh.2020.567885) [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Kaewsrichan J, Peeyananjarassri K & Kongprasertkit J. 2006. Selection and identification of anaerobic lactobacilli producing inhibitory compounds against vaginal pathogens. FEMS Immunol Med Microbiol 48 75–83. ( 10.1111/j.1574-695x.2006.00124.x) [DOI] [PubMed] [Google Scholar]
  21. Ketchem R, Rix M & Duttlinger V. 2018. Here’s How the U.S. Differs from Canada in Culling Sows. National Hog Farmer. (https://www.nationalhogfarmer.com/hog-reproduction/here-s-how-the-u-s-differs-from-canada-in-culling-sows) [Google Scholar]
  22. Kiefer ZE, Koester LR, Showman L, et al. 2021. Vaginal microbiome and serum metabolite differences in late gestation commercial sows at risk for pelvic organ prolapse. Sci Rep 11 6189. ( 10.1038/s41598-021-85367-3) [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Knol EF, Nielsen B & Knap PW. 2016. Genomic selection in commercial pig breeding. Anim Front 6 15–22. ( 10.2527/af.2016-0003) [DOI] [Google Scholar]
  24. Kraeling RR & Webel SK. 2015. Current strategies for reproductive management of gilts and sows in North America. J Anim Sci Biotechnol 6 3. ( 10.1186/2049-1891-6-3) [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Lillehammer M, Meuwissen THE & Sonesson AK. 2011. Genomic selection for maternal traits in pigs. J Anim Sci 89 3908–3916. ( 10.2527/jas.2011-4044) [DOI] [PubMed] [Google Scholar]
  26. Linhares IM, Summers PR, Larsen B, et al. 2011. Contemporary perspectives on vaginal pH and lactobacilli. Am J Obstet Gynecol 204 e1–e120. ( 10.1016/j.ajog.2010.07.010) [DOI] [PubMed] [Google Scholar]
  27. Lukac D, Vidovic V, Vasiljevic T, et al. 2016. Estimation of genetic parameters and breeding values for litter size in the first three parity of Landrace sows. Biotechnol Anim Husb 32 261–269. ( 10.2298/bah1603261l) [DOI] [Google Scholar]
  28. Luque AT, Fontana C, Pasteris SE, et al. 2021. Vaginal bacterial diversity from healthy gilts and pregnant sows subjected to natural mating or artificial insemination. Res Vet Sci 140 26–37. ( 10.1016/j.rvsc.2021.07.023) [DOI] [PubMed] [Google Scholar]
  29. Marrazzo JM 2011. Interpreting the epidemiology and natural history of bacterial vaginosis: are we still confused? Anaerobe 17 186–190. ( 10.1016/j.anaerobe.2011.03.016) [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Mclaren MR & Callahan BJ. 2021. Silva 138.1 prokaryotic SSU taxonomic training data formatted for DADA2 [Dataset]. Zenodo. (https://zenodo.org/records/4587955) [Google Scholar]
  31. Mcmurdie PJ & Holmes S. 2013. Phyloseq: an R package for reproducible interactive analysis and graphics of microbiome census data. PLoS One 8 e61217. ( 10.1371/journal.pone.0061217) [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Mirmonsef P, Zariffard MR, Gilbert D, et al. 2011. Short-chain fatty acids induce pro-inflammatory cytokine production alone and in combination with toll-like receptor ligands. Am Soc Reprod Immunol 67 391–400. ( 10.1111/j.1600-0897.2011.01089.x) [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Mirmonsef P, Gilbert D, Veazey RS, et al. 2012. A comparison of lower genital tract glycogen and lactic acid levels in women and macaques: implications for HIV and SIV susceptibility. AIDS Res Hum Retroviruses 28 76–81. ( 10.1089/aid.2011.0071) [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Monteiro MS, Poor AP, Muro BBD, et al. 2022. The sow microbiome: current and future perspectives to maximize the productivity in swine herds. J Swine Health Prod 30 238–250. ( 10.54846/jshap/1277) [DOI] [Google Scholar]
  35. Moreno LZ, Matajira CEC, Gomes VTM, et al. 2016. Molecular and antimicrobial susceptibility profiling of atypical streptococcus species from porcine clinical specimens. Infect Genet Evol 44 376–381. ( 10.1016/j.meegid.2016.07.045) [DOI] [PubMed] [Google Scholar]
  36. Murase K, Watanabe T, Arai S, et al. 2019. Characterization of pig saliva as the major natural habitat of Streptococcus suis by analyzing oral, fecal, vaginal, and environmental microbiota. PLoS One 14 e0215983. ( 10.1371/journal.pone.0215983) [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Oksanen J, Simpson G, Blanchet F, et al. 2022. _vegan: community Ecology Package_. R package version 2.6-4. [Google Scholar]
  38. Patel N, Patel N, Pal S, et al. 2022. Distinct gut and vaginal microbiota profile in women with recurrent implantation failure and unexplained infertility. BMC Womens Health 22 113. ( 10.1186/s12905-022-01681-6) [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Patterson J & Foxcroft G. 2019. Gilt management for fertility and longevity. Animals 9 434. ( 10.3390/ani9070434) [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Pedregosa F, Varoquaux G, Gramfort A, et al. 2011. Scikit-learn: machine learning in Python. JMLR 12 2825–2830. (https://jmlr.org/papers/v12/pedregosa11a.html) [Google Scholar]
  41. Poor AP, Moreno LZ, Monteiro MS, et al. 2022. Vaginal microbiota signatures in healthy and purulent vulvar discharge sows. Sci Rep 12 9106. ( 10.1038/s41598-022-13090-8) [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Popovac M, Radojković D, Petrović M, et al. 2012. Heritability and connections of sow fertility traits. Biotechnol Anim Husb 28 469–475. ( 10.2298/bah1203469p) [DOI] [Google Scholar]
  43. Quast C, Pruesse E, Yilmaz P, et al. 2013. The SILVA robosomal RNA gene databse project: improved data processing and web-based tools. Nucl Acids Res 41 D590–D596. ( 10.1093/nar/gks1219) [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. R Core Team 2023. R: a language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing. [Google Scholar]
  45. Ravel J, Gajer P, Abdo Z, et al. 2011. Vaginal microbiome of reproductive-age women. Proc Natl Acad Sci U S A 108 (Supplement 1) 4680–4687. ( 10.1073/pnas.1002611107) [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Renzhammer R, Loncaric I, Ladstätter M, et al. 2020. Detection of various streptococcus spp. and their antimicrobial resistance patterns in clinical specimens from Austrian swine stocks. Antibiotics 9 893. ( 10.3390/antibiotics9120893) [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Sanglard LP, Schmitz-Esser S, Gray KA, et al. 2020. Vaginal microbiota diverges in sows with low and high reproductive performance after porcine reproductive and respiratory syndrome vaccination. Sci Rep 10 3046. ( 10.1038/s41598-020-59955-8) [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Shannon CE & Weaver W. 1949. The Mathematical Theory of Communication, pp 1–125. Urbana, Illinois: University of Illinois Press. [Google Scholar]
  49. Simpson EH 1949. Measurement of diversity. Nature 163 688. ( 10.1038/163688a0) [DOI] [Google Scholar]
  50. Soede NM, Langendijk P & Kemp B. 2011. Reproductive cycles in pigs. Anim Reprod Sci 124 251–258. ( 10.1016/j.anireprosci.2011.02.025) [DOI] [PubMed] [Google Scholar]
  51. Sorenson T 1948. A method of establishing groups of equal amplitude in plant sociology based on similarity of species content. Kongelige Danske Videnskabernes Selskab 5 1–34. [Google Scholar]
  52. Tachedjian G, Aldunate M, Bradshaw CS, et al. 2017. The role of lactic acid production by probiotic Lactobacillus species in vaginal health. Res Microbiol 168 782–792. ( 10.1016/j.resmic.2017.04.001) [DOI] [PubMed] [Google Scholar]
  53. Wang J, Li C, Nesengani LT, et al. 2017. Characterization of vaginal microbiota of endometritis and healthy sows using high-throughput pyrosequencing of 16S rRNA gene. Microb Pathog 111 325–330. ( 10.1016/j.micpath.2017.08.030) [DOI] [PubMed] [Google Scholar]
  54. Wang C, Fan A, Li H, et al. 2020. Vaginal bacterial profiles of aerobic vaginitis: a case–control study. Diagn Microbiol Infect Dis 96 114981. ( 10.1016/j.diagmicrobio.2019.114981) [DOI] [PubMed] [Google Scholar]
  55. Wang M, Yue J, Lv G, et al. 2024. Effects of interactions between feeding patterns and the gut microbiota on pig reproductive performance. Animals 14 2714. ( 10.3390/ani14182714) [DOI] [PMC free article] [PubMed] [Google Scholar]
  56. Wright ES 2016. Using DECIPHER v2.0 to analyze big biological sequence data in R. R Journal 8 352–359. ( 10.32614/rj-2016-025) [DOI] [Google Scholar]
  57. Zhang J, Liu M, Ke S, et al. 2021a. Gut and vagina microbiota associated with estrus return of weaning sows and its correlation with the changes in serum metabolites. Front Microbiol 12 690091. ( 10.3389/fmicb.2021.690091) [DOI] [PMC free article] [PubMed] [Google Scholar]
  58. Zhang L, Wang L, Dai Y, et al. 2021b. Effect of sow intestinal flora on the formation of endometritis. Front Vet Sci 8 663956. ( 10.3389/fvets.2021.663956) [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Data Availability Statement

All relevant sequencing data included in this study is available in the Agri-environmental Research Data Repository at the University of Guelph: https://doi.org/10.5683/SP3/C6Y2LN.

Articles from Reproduction (Cambridge, England) are provided here courtesy of Bioscientifica Ltd.

RESOURCES