Abstract
A non-optimal vaginal microbiome (VMB) is typically diverse with a paucity of Lactobacillus crispatus and is often associated with bacterial vaginosis (BV) and sexually transmitted infections (STIs). Although compositional characterization of the VMB is well-characterized, especially for BV, knowledge remains limited on how different groups of bacteria relate to incident STIs, especially among adolescents. In this study, we compared the VMB (measured via 16S ribosomal RNA gene amplicon sequencing) of Kenyan secondary school girls with incident STIs (composite of chlamydia, gonorrhea, and trichomoniasis) to those who remained persistently negative for STIs and BV over 30 months of follow-up. We applied microbial network analysis to identify key taxa (i.e., those with the greatest connectedness in terms of linkages to other taxa), as measured by betweenness and eigenvector centralities, and sub-groups of clustered taxa. VMB networks of those who remained persistently negative reflected greater connectedness compared to the VMB from participants with STI. Taxa with the highest centralities were not correlated with relative abundance and differed between those with and without STI. Subject-level analyses indicated that sociodemographic (e.g., age and socioeconomic status) and behavioral (e.g., sexual activity) factors contribute to microbial network structure and may be of relevance when designing interventions to improve VMB health.