Skip to main content

Poster Session 2025

Defining the female genital tract virome

Presented By: Askarbek Orakov

The female genital tract (FGT) microbiome is highly dynamic, with a unique ecology that exists as distinct community state types across time and between women. Compositional differences in the FGT microbiome, such as variation in Lactobacillus species and strains, are strongly associated with women’s and infant health, ranging from risk of sexually transmitted infections to vaginosis, and preterm delivery.  Like most environments, the viral composition of the FGT has been understudied relative to non-viral microbes, due to the great diversity of viral biology, their intractability in most experimental settings, and limited computational and experimental methods. However, both eukaryotic viruses and phage in this environment are critical for host and bacterial interactions that determine immune and health phenotypes. We thus provide the first comprehensive characterization of the FGT virome integrating 1) two US cohorts with paired, longitudinal metagenomic and metatranscriptomic FGT samples, totalling 80 subjects, and 2) non-Western cohorts from Bangladesh and Zambia with 412 and 291 metagenomic FGT samples, respectively. All samples were subjected to a new approach to viral profiling, BAQLaVa, which enables dramatically more sensitive and specific viral detection by performing nucleotide and translated searches against an integrated database of taxon-specific markers. We separately analyzed the viral and non-viral domains of these FGT microbiomes and discovered viral community structures that both interact with and are distinct from those of corresponding non-viral microbes. We also identified geographical enrichment of viral taxa and ecologies, as well as systematically enumerating putative phage-host relationships, leading to novel insights into phage predation in FGT microbiomes. In the next steps, we aim to identify greater functional and community dynamics insights from these data by integration of paired metagenomic/metatranscriptomic profiles and from longitudinal modeling, respectively.