Significance of a diagnostic model for detecting reproductive: tract microbiota profiles associated with female infertility using metagenomic approaches

Abstract

This study profiled the vaginal and endometrial microbiota of infertile women (n = 30) and fertile controls (n = 30) via shotgun metagenomic sequencing. Raw reads were processed through the Fastp → HUMAnN 4 → MetaPhlAn 4 pipeline; α-diversity (Shannon), β-diversity (Bray–Curtis, PERMANOVA) and differential taxa (LEfSe, LDA > 2) were assessed. A Random Forest classifier built on the 15 most informative taxa achieved an AUC of 0.85, sensitivity 0.88 and specificity 0.82. Findings demonstrate a strong association between CST-IV polymicrobial dysbiosis—particularly the Gardnerella vaginalis + Prevotella bivia signature—and infertility, whereas depletion of Lactobacillus crispatus/jensenii reflects loss of a protective niche. The microbiome-based model provides sufficient accuracy for clinical screening and underpins development of novel probiotic–prebiotic therapeutic strategies.