Vaginal Microbiome in Pregnancy and HIV

16S-based meta-analysis of the vaginal microbiome in pregnant people living with HIV.

Overview

A pilot study and global meta-analysis of the vaginal microbiome in pregnant people living with HIV on antiretroviral therapy, examining community state types, host immune mediators, and associations with adverse reproductive outcomes.

Scientific or practical problem

The vaginal microbiome is linked to reproductive health outcomes, but data integration across studies of pregnant people living with HIV has been limited. This project combined a pilot cohort with a broader meta-analysis to look for consistent microbiome–immune associations.

My role

Statistical analysis. I designed and executed the statistical analysis plan for a human cohort (82 pregnant people living with HIV) together with a global meta-analysis (1,800+ samples), applying compositional data analysis, regression, non-parametric testing, batch correction, and differential abundance analysis.

Dataset or data type

16S rRNA amplicon sequencing data from a pilot cohort (Democratic Republic of Congo) integrated with publicly available vaginal microbiome datasets for meta-analysis, alongside host immune mediator measurements.

Methods and technologies

• 16S amplicon analysis and functional prediction
• Compositional data analysis
• Batch correction across integrated datasets
• Regression and non-parametric statistical testing
• Differential abundance analysis
• Integration of microbiome and host immune-mediator data

Workflow diagram

flowchart LR
  A[Pilot cohort 16S data] --> C[Batch-corrected integration]
  B[Public meta-analysis datasets] --> C
  C --> D[Compositional / differential abundance analysis]
  D --> E[Microbiome-immune association testing]

Major results or outputs

Published pilot study and meta-analysis identifying associations between vaginal community state types, inflammatory immune mediators, and reproductive outcomes in this population.

Challenges and decisions

Integrating multiple independently generated 16S datasets required batch-effect correction before any cross-cohort comparison could be trusted — a key methodological decision that shaped the whole analysis plan.

Publications

Ndlovu, K.S., Pavan, R.R., Corry, J., Gregory, A.C., Mahamed, S., Zotova, N., Tabala, M., Babakazo, P., Funderburg, N.T., Yotebieng, M. and Klatt, N.R., 2026. The vaginal microbiome of pregnant people living with HIV on antiretroviral therapy in the Democratic Republic of Congo: a pilot study and global meta-analysis. mSphere, e00597-25. https://doi.org/10.1128/msphere.00597-25

Code and documentation

Analysis code repository link pending.

Collaborators

Collaborative study led by K.S. Ndlovu and colleagues (Klatt Lab); statistical analysis contributed by R.R. Pavan. Participant-level information is not disclosed, consistent with the published study’s ethical approvals.

Current status

Published (2026).

Limitations or confidentiality note

No participant-identifying information is presented here; all details reflect what is reported in the published paper.