Predicting bacterial vaginosis incidence using artificial neural networks

The useful part

George, Megan Amerson-Brown, Meng Luo, Ashutosh Tamhane, Paweł Łaniewski, Alison J. Taylor Background Bacterial vaginosis (BV) is a vaginal dysbiosis associated with adverse reproductive and infectious outcomes. We evaluated whether artificial neural network (ANN) models of the vaginal microbiome could predict incident BV (iBV) up to 14 days before clinical diagnosis.

How it works

• Interpretation Vaginal microbiome composition contains predictive information that can identify women at risk of iBV before clinical onset.
• SHAP analysis was used to identify taxa associated with model predictions.
• as the taxa most strongly associated with model predictions.
• Models classified individual specimens as pre-iBV or healthy using the relative abundance of vaginal bacterial taxa.
• Model performance was assessed using a held-out participant-level test set, participant-level cross-validation, and external validation.

What to take from it

Results On the held-out participant-level test set, the ANN achieved 93% accuracy (AUC = 0.97, sensitivity = 95%, specificity = 92%). mulieris, and Megamonas spp.) maintained >91% accuracy, sensitivity, and specificity.

Details worth keeping

Current diagnostics identify BV after symptom onset. Models using only five taxa (Lactobacillus crispatus, Gardnerella spp., L.