Related Experiment Video
Updated: Jul 3, 2025
08:15
Author Spotlight: Revolutionizing Research on Vaginal Microbiome Interactions Using a Vaginal Chip
Published on: February 16, 2024
2.3K
DeepMPTB: a vaginal microbiome-based deep neural network as artificial intelligence strategy for efficient preterm
Oshma Chakoory1, Vincent Barra2, Emmanuelle Rochette3
1Université Clermont Auvergne, INRAE, MEDIS, F-63000, Clermont-Ferrand, France.
Biomarker Research
|February 14, 2024
Summary
Predicting preterm birth (PTB) risk is crucial for neonatal health. A deep neural network analyzing vaginal metagenomics identified overall microbial diversity, not specific species, as key for accurate PTB prediction.
Area of Science:
- Microbiome research
- Computational biology
- Maternal-fetal medicine
Background:
- Preterm birth (PTB) is a leading cause of neonatal mortality globally.
- Vaginal microbiome structure is increasingly recognized for its role in pregnancy outcomes.
- Accurate prediction of PTB remains a significant clinical challenge.
Discussion:
- A deep neural network (DNN) was developed and trained using vaginal metagenomics data from 561 pregnant women.
- The DNN model achieved 84.10% accuracy and an AUROC of 0.875 ± 0.11 in predicting term birth (TB) and PTB.
- Benchmarking confirmed the DNN model's superior performance compared to seven existing machine learning algorithms.
Key Insights:
- Overall vaginal microbial diversity is a more significant predictor of PTB than specific microbial species.
- The developed artificial intelligence (AI) strategy offers a novel approach for PTB risk assessment.
- This AI-driven tool can aid clinicians in providing personalized care for pregnant individuals.
Outlook:
- Further validation of the AI model in diverse populations is warranted.
- Integration of this predictive tool into routine clinical practice could improve neonatal outcomes.
- Continued research into the complex interplay between the microbiome and PTB is essential.
