Figure 1.
Pie charts representing the bacterial distribution found in each of the samples represented.
Subject 101 is representative of 7 other patients with similar profiles (Subjects 104, 105, 106, 107, 109, 113 and 114). Subject 103 and Subject 110 have similar profiles. Subject 108 and Subject 111 have similar profiles. Divisions within the pie charts represent different OTUs (Operational Taxonomic Units – Strain level). “Other” represents OTUs with ≤1% abundance that belong to various species, excluding L. crispatus and L. iners.
Figure 2.
Principal Component Analysis (PCA) of temporal dynamics of the vaginal microbiome during pregnancy.
Arrows indicate shifts from the first timepoint, taken at 8–12 weeks, and the subsequent timepoints. Notice that all temporal profiles that involve L. iners dominance involve a shift in the microbiome profile while purely L. crispatus dominated temporal profiles do not. Visualization carried out using Matplotlib version 1.2.1 (The MathWorks, Inc.).
Table 1.
Significant parameters found in the metadata.
Figure 3.
Shannon Diversity Index for the vaginal microbiome samples.
The profile characterized by a dominance of L. crispatus throughout pregnancy (shared by 8 of the subjects) is significantly less diverse than the two other profiles (except during the first trimester). * p<0.05; ** p<0.005 on paired t-test.
Figure 4.
Pregnancy effect in African-American and Caucasian subjects as measured by Chao1 diversity Index.
Diversity is significantly reduced during pregnancy in both ethnicities (**p<0.01, Monte Carlo analyses, 999 permutations).
Figure 5.
Unweighted Unifrac beta-diversity between pregnant and non-pregnant subjects within each ethnicity.
Caucasians show significant convergence as a result of pregnancy (results confounded by platform effects – See Table S1) while African-Americans show significant divergence in pregnancy when compared to non-pregnant subjects. **p<0.001, Monte Carlo analyses, 999 permutations.
Figure 6.
Lactobacillus species frequencies across ethnicities and pregnancy states.
Significant differences were found between all categories except between pregnant Caucasians and pregnant African-Americans (*p<0.05;**p<0.01; Monte Carlo analysis, 999 permutations).
Figure 7.
Estimation of species richness according to Chao 1 index.
A – African-American subjects (Red – N009, Orange – N018, Blue – N017). Diversity between subjects is significant and therefore does not allow for the grouping of the subjects for analytical purposes. B – Caucasian subjects. No significant differences in diversity were found among subjects.
Figure 8.
Unweighted PCA analysis by Subject and Trimester.
A – African-American by Subject. Clustering by Subject can be observed. B – Caucasian by Subject. No clear clustering is observed. C – African-American by Trimester. No common attractor can be identified. D – Caucasian by Trimester. A common attractor can be observed. Plots C and D – 1st Trimester – Red; 2nd Trimester – Blue; 3rd Trimester – Orange.