3.1 Clinical Characteristics of the Study Subjects
The demographic characteristics of the study participants in the stress urinary incontinence (SUI) group (case group, n=13) and Non-SUI group (control group, n=19) were presented in Supplementary Table S1. No significant differences were observed between the case group and the control group in terms of age, gestation, partitioning, weight gain, and birthweight. However, it was worth noting that a marginal difference in BMI was detected between the two groups (P = 0.03).
3.2 The Diversity of the Vaginal Microbiota
Dilution curves were computed and recorded after five random samples, with a minimum sampling depth of 50,000. As sampling depth increased, dilution curves leveled off, indicating reasonable sequencing data coverage. (Supplementary Figure S1). For the alpha diversity indexes calculated in this study, the Shannon-Weaver index (P < 0.05) (Figure 1A), Gini-Simpson index (P < 0.01) (Figure 1B) and Pielou (Figure 1C) were significantly higher in group SUI compared to group Non-SUI, indicating that the vaginal microbial community composition of patients with SUI was more complex and diverse. The NMDS analysis effectively simulates microbial communities, with differences in vaginal microbiota between groups and stress = 0.191 (< 0.2) (Figure 1D, E). The beta diversity of the two groups was compared by primary coordinate analysis (PCoA) and ANOSIM test (R = -0.02, P = 0.634) ((Figure 1F and Supplementary Figure S2). The PCoA results showed that the difference between groups was greater than that within groups, implying a significant difference in diversity among group SUI and Non-SUI.
3.3 The Composition and Biomarkers of the Vaginal Microbiota
There were commonalities and differences in bacterial composition between group SUI and Non-SUI. At the phylum level, the microbial community composition of both groups was dominated by Firmicutes, Actinobacteriota, Bacteroidota, and Proteobacteria, but the abundance percentage of the components varied at the phylum level, especially Bacteroidota (Figure 2A). At the genus level, most of the samples in Non-SUI group were still dominated by Lactobacillus (57.9%). In addition, in the remaining samples from the Non-SUI group, the bacteria with higher abundance were Prevotella, Gardnerella, Streptococcus. But in SUI group, the microbial community composition of most samples was no longer dominated by Lactobacillus but showed increased abundance of other bacteria such as Prevotella, Gardnerella, and Streptococcus. The overall microbial community composition was more complex in SUI group. Microbial dysbiosis was observed [23, 24] (Figure 2B). Based on the proportion of samples with dysbiosis of the vaginal microbiota in two groups (whether the vaginal flora is dominated by Lactobacillus). The statistical results showed 42.1% (n = 8) of the Non-SUI group (n = 19) and 84.6% (n = 11) of the SUI group (n = 13) had dysbiosis (Figure 2C). The results of logistic regression analysis showed a statistically significant effect of SUI on vaginal microbial dysbiosis (P = 0.016, OR = 82.977) and the delivery method of patient was a key factor contributing to dysbiosis (OR=1.358) (Supplementary Table S2). The upset analysis revealed the presence of certain bacterial genera in the samples. Specifically, Lactobacillus, Prevotella, and Streptococcus were present in all samples, indicating their high abundance across the dataset. Additionally, the Gardnerella was detected in 31 samples, suggesting its relatively common occurrence. The Anaerococcus was observed in 23 samples, indicating its presence in a substantial number of samples as well (Figure 2D).
3.4 Biomarkers of Sample Groups Discovered by LEfSe, DESeq2, and Metastats
To better elucidate the vaginal microbiota biomarkers among the Non-SUI and SUI groups, three methods were used to analyze the markers. We conducted an analysis on the variation in bacterial abundance between the Non-SUI and SUI groups and visualized the differences using a Manhattan plot. The top 10 bacteria with the most significant abundance variations between the two groups were labeled on the plot. Compared to the Non-SUI group, the SUI group exhibited an elevated abundance of low-abundance bacteria, such as Streptococcus, Prevotella, Dialister, and Veillonella, and a reduction in the abundance of Lactobacillus (Figure 3A). Volcano plot showed that Gardnerella, Prevotella, and Dialister in group SUI were biomarkers ((log2FoldChange > = 2, P < 0.05) (Figure 3B). At the phylum level, the distribution of the phylum Bacteroidota was significantly different between the two groups. Group SUI had a significantly higher abundance of Bacteroidota (P < 0.05), while Firmicutes were more abundant in group Non-SUI (Figure 3C). Further analysis of bacterial abundance differences at the phylum level showed the same results as manhattan. Among the top 15 genera in abundance, the distribution of Lactobacillus, Streptococcus, Prevotella, Finegoldia and Dialister was significantly different between the two groups (P < 0.05). Lactobacillus exhibited significant up-regulation in the Non-SUI group, while Streptococcus, Prevotella, Finegoldia, and Dialister were significantly up-regulated in the SUI group. Additionally, the average abundance of Gardnerella was higher in the SUI group compared to the Non-SUI group (Figure 3D).
The results of LEfSe showed that the identified intergroup biomarker genera echoed the differences previously demonstrated in the manhattan plots at the genus level (LDA > 4, P < 0.05). The results indicated that Lactobacillus, Prevotella, Streptococcus, Finegoldia, Dialister, Hirschia, and Megasphaera were intergroup differential biomarker genera (Figure 4A, B). In addition, we compared the differences in the distribution of the four differentially marker bacteria with higher abundance in groups Non-SUI and SUI. Compared to group Non-SUI, group SUI showed a lower average abundance of Lactobacillus. The higher abundance of Lactobacillus in most samples in group Non-SUI, while only two samples in group SUI contained high abundance of Lactobacillus. Prevotella, Streptococcus and Dialister were more frequent and of higher average abundance in group SUI (Figure 4C).
3.5 Network Co-occurrence Analysis and Interactions among Vaginal Microbiota
To unravel the relationship between microorganisms, we performed network co-occurrence analysis and mental test. With the same network construction parameters (correlation r > 0.6 or r < – 0.6, P < 0.05), the network in group Non-SUI had 96 nodes and 133 edges (Figure 5A, Supplementary Table S3) while group SUI had 200 nodes and 409 edges (Figure 5B). Next, network attribute analysis was performed, and the average degree was 2.771 for group Non-SUI and 4.07 for group SUI. The average degree of group SUI was significantly higher than that of group Non-SUI (P < 0.0001) (Figure 5C). The number of triangles, the number of sides forming triangles was significantly higher in group SUI than in group Non-SUI (P < 0.001) (Figure 5D). The average clustering coefficient of the SUI was significantly higher than that of Non-SUI (P < 0.01) (Figure 5E). Moreover, we conducted a mental test to examine the interactions among the top 10 most abundant bacterial genera in the Non-SUI and SUI groups, respectively. It performed a Mental test of their interactions, which showed that Gardnerella and Lactobacillus in group Non-SUI showed a significant negative correlation (P < 0.01) and a significant positive correlation with Aerococcus (P < 0.01), Veillonella, Anaerococcus and Dialister showed a significant positive correlation (P < 0.001). Additionally, Dialister, Prevotella, and Anaerococcus showed a significant positive correlation (P < 0.01, P < 0.001), Anaerococcus and Prevotella showed significant positive correlation (P < 0.001) (Figure 5F). In SUI group, there is a significant positive correlation between Dialister and Anaerococcus (P < 0.01), as well as between Finegoldia and Corynebacterium (P < 0.01) (Figure 5G).