Clinical and Vaginal Microbiome Data from 43 Pregnant Women in Mexico City

A Mexican cohort study provides longitudinal vaginal microbiome and clinical data from 43 pregnant women (110 samples, 14 preterm births) recruited from public hospitals in Mexico City. The dataset includes genus-level 16S rRNA sequencing profiles and clinical variables, analyzed using leakage-aware machine learning frameworks. It was published by Martín Ruhle on figshare under a CC-BY-4.0 license in May 2026.

Use Cases

• Training descriptive machine learning models based on combined clinical and microbiome features.
• Analyzing associations between vaginal microbiome taxa (e.g., Mycoplasma, Peptostreptococcus) and preterm birth risk.
• Evaluating feature stability and importance for anthropometric variables like BMI and pre-pregnancy weight.
• Benchmarking leakage-aware analytical frameworks for small cohort studies in clinical microbiology.

Strengths

• Includes 110 longitudinal samples from 43 subjects, providing repeated measures.
• Applies a rigorous nested cross-validation design to prevent data leakage.
• Reports specific model performance metrics (e.g., AUROC 0.813±0.110) and differential abundance results.

Limitations

• Small sample size (43 women, 14 preterm births) limits statistical power and generalizability.
• Column-level documentation is absent; field semantics must be inferred from the accompanying PDF.
• The 312.1 KB size suggests the primary data is a summary document, not the raw underlying dataset.

Provenance

Source

figshare, authored by Martín Ruhle.

Collection Method

Vaginal microbiome profiles from genus-level 16S rRNA V3-V4 sequencing of samples from pregnant women.

Freshness

Last updated 2026-05-07 05:43:03; freshness should be verified.

Geography

Participants recruited from public hospitals in Mexico City, Mexico.