Page 139 - Read Online

Linehan et al. Microbiome Res Rep. 2025;4:24 https://dx.doi.org/10.20517/mrr.2024.92 Page 5 of 20

associated with low DNA concentration or negative controls [25,27,28] . Samples with < 5 k reads and genera
present in ≤ 10% of samples were excluded. A phylogenetic tree was built using DECIPHER (v2.16.1) and
phangorn (v2.5.5) [29,30] . Functional predictions were made with PICRUSt2 (v2.5.1) against the IMG database,
and abundant KEGG pathways (Level 3) were identified [31-33] .

Statistical analysis
Analyses were conducted in R (v4.1.2) with visualization via ggplot2 (v3.4.1). Core ASVs were determined
by presence in ≥ 50% of samples and abundance > 0.001 . Annotation was further refined using BLASTn
[34]
[35]
against the NCBI 16S rRNA database . Five perinatal factors (maternal antibiotic use, delivery mode,
feeding type, infant gender, and PROM) were assessed for their impact on the microbiome. Alpha diversity
(Chao1, Shannon, Simpson) was calculated using the iNEXT package and differences analyzed via linear
[36]
models. Principal component analysis (PCA) was performed on clr-transformed values for beta diversity,
[37]
with zero imputation using the “const” method . Aitchison distances were calculated for beta diversity,

and differences were analyzed using PERMANOVA (adonis function, vegan package v2.6.4), with
Bonferroni-Holm correction. Significance was set at P < 0.05. Differentially abundant taxa were identified
[39]
[38]
using ANCOM-BC , while KEGG pathway differences were detected with STAMP (v2.1.3) .
RESULTS
The placental does not contain a discernable microbiota
Sequencing of placental tissues produced 666,018 reads (mean length 251 bp), with an average of 79,820
reads per sample (SD ± 111,793). Post-quality filtering, dereplication, error modeling, denoising, pair
merging, and chimera removal (using DADA2 default parameters), 10 samples contained zero reads and
were excluded, leaving 8 samples with an average of 29,774 reads (range 8,630-38,251). Out of 964 ASVs
detected across the 10 samples, 835 were identified as contaminants by Decontam, representing 43.7%-67%
(median 57%) of ASVs per sample. The remaining 129 ASVs could not be taxonomically resolved to the
species level. Filtering steps applied to other sample types, which excluded genera present in ≤ 10% of total
samples, resulted in no ASVs for placental samples; this step was thus removed for placental analysis. DNA
extraction and amplicon PCR blanks were used for comparison. Actinobacteriota, Firmicutes, and
Proteobacteria were the only phyla detected across blanks and placental samples [Supplementary Figure 1A
]. Five families (Bacillaceae, Corynebacteriaceae, Micrococcaceae, Streptococcaceae, Xanthobacteraceae) and
six genera (Afipia, Bacillus, Corynebacterium, Enhydrobacter, Micrococcus, Streptococcus) were identified [
Supplementary Figure 1B and C].

Site-specific shared microbial taxa between mother and infant
An overview of the phylogenetic diversity, core microbiome and metabolic pathways in infant and maternal
samples are shown in Supplementary Figures 2-5. To explore site-specific shared bacterial taxa of maternal
microbiota species to the infant, we identified bacterial taxa shared between the communities of mothers
and their related infants. Figure 1 gives an overview of the overall microbial composition and the relative
abundances of the four sample types investigated in this study. Regarding the establishment of the infant’s
oral microbiome, on average, a related mother and infant’s oral microbiota shared 45 taxa [Figure 2A],
accounting for 65% of the total reads in the infant’s sample [Figure 2B]. Perinatal factors had no effect on
the amount of the infants’ oral microbiome that was shared with their mothers. Rothia mucilaginosa was
found present in all 18 mother-infant oral sample dyads. Streptococcus oralis, Haemophilus parainfluenzae
and Fusobacterium nucleatum were found in 15 dyads. On average, a related mother’s vaginal microbiome
and the infant’s oral microbiota shared 15 bacterial taxa [Figure 2C], which accounted for 15% of the total
reads in the infant’s sample [Figure 2D]. When comparing natural birth (NB) infants with those born by CS,
we found that the mode of delivery had a significant effect on the amount of the infants’ oral microbiome
that was shared with their mother’s vaginal microbiome (Mann Whitney Test, P = 0.045) [Supplementary