Prevotella stercorea defines dual ecological pathways linking the gut microbiome to infection susceptibility. Ogochukwu Ofordile

Research hypothesis
We hypothesised that Prevotella stercorea, previously shown to associate with reduced infection in Gambian children, exerts protection through two mechanistically distinct ecological pathways: a community-mediated pathway operating through microbiome richness and colonisation resistance, and a species-autonomous pathway independent of community structure. We further hypothesised that the species-autonomous pathway would be reflected in immune activation conditioned on host nutritional status.
What the data shows
This dataset contains microbiome taxonomic abundance data, clinical metadata, and adverse event records from 633 children aged 6–35 months enrolled in the IHAT-GUT randomised controlled trial (NCT02941081), conducted in the Upper River Region, The Gambia. Stool samples were collected at Days 1, 15, and 85 of follow-up. Adverse events including acute respiratory infection (ARI), diarrhoea, and fever were recorded over a 113-day follow-up period. Systemic inflammatory biomarkers (CRP, AGP, calprotectin) and anthropometric measures are included.
Notable findings
P. stercorea abundance positively predicts community species richness, consistent with a role in community assembly. Its association with reduced ARI (IRR = 0.946, p = 0.002) is unchanged after richness adjustment and richness itself is not associated with ARI, satisfying the criterion for a richness-independent pathway. A co-occurring congener, P. copri, does not retain an independent ARI association in joint models. The inflammatory signature of P. stercorea is detectable only in better-nourished children, consistent with immune activation blunted by nutritional anergy.
Data interpretation and use
Species-level microbiome abundances are expressed as relative abundances and should be log-transformed (log₁+x) prior to regression modelling. Richness is defined as the count of taxa with non-zero abundance per sample. Adverse event frequencies and durations are capped at 5 episodes and 30 days respectively to limit outlier influence. The complete reproducible analysis pipeline, including all R scripts and model specifications, is available at https://github.com/ofordile-star/IHAT_Pstercorea.