In the Bray–Curtis PCoA plots we observed (a) separation of the Canadian SA cohort from both the Indian and American cohorts. The open circles represent samples from GEN1 moderately recent (mid) and early immigrant groups. Boxplots show the dissimilarity distances for each group, on each axis. (b) This separation was primarily driven by the maximum taxa, with P. copri, Bacteroides spp., or D. succinatiphilus dominating the majority of the communities in the Indian cohort, and Bacteroides spp. or Clostridia spp. dominating the majority of communities in the American cohort. (c) A significantly higher percent of energy was derived from carbohydrates compared to proteins in the GEN1 recent immigrant group, compared to other YAI groups and GEN2, as determined by the food frequency questionnaire (FFQ). Participant macronutrient data was not available (NA) for external data sets. (d) D. succinatiphilus was significantly enriched (log-fold change (LFC) > 3, p < .001) in the Indian cohort and recent SA Canadian immigrants, compared to the early immigrants and GEN2. D. invisus is significantly depleted, particularly for the residents of Bhopal, who consumed primarily a plant-based diet (LFC < −1, p < .001). We extracted all the species of (e) Veillonellaceae (class Negativicutes), (f) Prevotella, and Bacteroides from the data to observe trends in the abundances of these taxa across all cohorts and YAI groups