The vaginal microbiome (VMB) plays an important role in women’s reproductive health. An optimal VMB is characterized by the dominance of Lactobacillus species, which maintain an acidic environment and inhibit pathogenic bacteria. A shift to a diverse anaerobic microbial community, including Gardnerella vaginalis and Prevotella species, is associated with bacterial vaginosis (BV), affecting nearly 30% of women in the United States and linked to adverse health issues. While antibiotics (metronidazole (MNZ) or clindamycin) are commonly used to treat BV, the long-term recurrence rate remains as high as 50%. Though BV recurrence has been linked to multiple physiologic and host lifestyle factors, none have been identified as the primary cause of treatment failure. Several alternative strategies, including biofilm removal, pH-lowering agents, and combinatorial regimens, have been employed to reduce recurrence rates, however, none have been successful. Adding to this complexity, the stability of the vaginal microbiome varies, with women experiencing both temporary and permanent shifts due to menstruation or treatment-related perturbations. Probiotics have recently been explored as an alternative treatment approach and may hold promise in improving BV outcomes. 

My research is leveraging the advantages of both mechanistic models and artificial intelligence (AI) to delineate novel microbial parameters, species interactions, and metabolic events linked to treatment efficacy against BV. This work will enable the identification of key biomarkers associated with BV transitions, as well as distinguish stable microbial compositions from unstable ones in response to menstruation or treatment interventions. Using mechanistic ordinary differential equations (ODE) models, I uncover new biological insights associated with treatment efficacy in BV+ patients, such as the importance of interaction terms vs. dosing, the influence of pretreatment composition, and the impact of non-target MNZ uptake. Next, I investigate the association between pre-menstruation microbial taxonomy and metabolomics with BV occurrence post-menses using AI for time-series analysis. I apply machine learning and deep learning algorithms to identify new species-metabolic interactions at pre-menses timepoints to predict BV after menses. My final objective is to design digital twin models, for the first time in VMB studies, to guide AI algorithms in the prognosis of BV+ patients. 

You can find my contributions to the field of Biomedical Engineering and women's health in my Google Scholar.