Objectives: to assess the relevance of the RiboTaxa algorithm coupled with neural network learning based on analysis of vaginal microbiota metagenomic sequencing data for predicting prematurity in an identified at-risk population.

Study description: Longitudinal follow-up of a cohort of pregnant women, with collection of biological samples, and a posteriori case-control comparison based on the occurrence of an event (premature birth).

There is currently no reliable clinical or biological diagnosis to predict premature birth. Recent work using metagenomic data analysis coupled with artificial intelligence approaches suggests that there may be a vaginal microbiota signature during pregnancy that correlates with the occurrence of preterm birth. The aim of the study is to use biological samples to confirm the identification of these vaginal microbiota signatures as a means of predicting preterm birth.

Inclusion Criteria:

• Pregnant women Admitted to the Clermont-Ferrand University Hospital maternity ward from 14 weeks' gestation onwards
• For threat of preterm birth (PTB) characterized by contractile activity and/or cervical changes, or for premature rupture of fetal membranes (PROM)
• And in need of vaginal swabbing
• Single or multiple pregnancy
• Able to understand and object to the study
• Covered by a French social security scheme.
• Give informed consent for the study

Exclusion Criteria:

• Patient under guardianship, curatorship or safeguard of justice
• Patient having received antibiotic therapy in the 2 weeks prior to admission