Integrating multi-omics data: Modeling biomarkers and mechanisms to reduce bacterial vaginosis recurrence

点击翻译按钮获取中文摘要

Bacterial vaginosis (BV) is characterized by a vaginal microbiota with a low abundance of Lactobacillus spp. and higher abundances of anaerobic bacteria and affects nearly 30% of reproductive-age North American women and closer to 50% of sub-Saharan African women. BV is diagnosed by clinical observation of Amsel’s criteria (Amsel-BV), but treatment is only recommended when symptoms are reported, leaving a significant proportion of women untreated. Even with treatment, BV recurrence rates range from 50-70% within 6 months, increasing a woman’s risk of negative sequelae. Regardless of symptoms, BV is associated with serious adverse health outcomes, including preterm birth and HIV, and can seriously impact a woman’s quality of life. Ideal BV treatment would eliminate recurrence. The vaginal microbiome and microenvironment together provide a detailed evaluation of BV states, and hold key functional insights to predict and understand Amsel-BV recurrence. The goal of this proposal is to integrate and operationalize microbiome (metagenomes) and microenvironment (metabolomes and immune markers) data to develop a prognostic indicator of recurrent BV, and identify candidate biomarkers and causal mechanisms which reduce recurrence. Recent work by the PI functionally categorized the vaginal microbiome for use in large clinical research studies (vaginal metagenomic community state types, mgCSTs). The broad hypothesis in this proposal is that not all microbiomes associated with bacterial vaginosis have the same potential for recurrence. Preliminary data suggest that BV recurrence is more frequently observed in only two of the nine mgCSTs containing BV-associated bacteria. This study proposes to utilize archived cervicovaginal samples from the NIH 1999 Longitudinal Study of Vaginal Flora in which participants were followed quarterly for one year. Multi-omic analyses of baseline samples will be assessed to identify microbial (metagenomic and metabolomic) and host (metabolomic and targeted immune markers) signatures of susceptibility to recurrent BV. Specific aims of this proposal are to: (1) conduct an epidemiological analysis to evaluate the demographic and lifestyle correlates of mgCSTs, and (2) employ supervised machine learning and causal inference modeling to identify prognostic factors and drivers of the vaginal microbiome and microenvironment which lead to recurrent BV. Cases are defined as women with Amsel-BV at baseline, then clearance 3 months later, followed by recurrence at six months. Controls are women with Amsel-BV that do not experience recurrence within 9 months. This grant will support the PI’s training in epidemiology and biostatistics with the completion of a Certificate Program in Clinical Research. The PI’s long- term goal is to create an independent research program translating the basic science of the vaginal microbiome to improve women’s reproductive health outcomes. The Institute for Genome Sciences and the University of Maryland School of Medicine uniquely provide the resources and support required for successful completion of this proposal and the PI’s transition to an independent investigator.