Fig. 2

Associations between maternal diet and infant gut microbial communities in infants delivered vaginally. Plots include only infants delivered by Cesarean section (n = 48). a Principal coordinate plot of generalized UniFrac distances (PERMANOVA p = 0.034 for maternal dairy as a continuous variable), colored by maternal dairy intake tertiles. Each point represents an individual, and lines indicate the distance from tertile centroid. b Predicted probability plot of infant stool cluster membership by maternal dairy intake during pregnancy from multinomial logistic regression models adjusted for infant feeding method, maternal BMI, parity, and batch. Cluster 1 is the reference group. c Linear model associations between maternal aMED score and relative abundance of infant stool OTUs. The size of each point indicates the log-ratio transformed relative abundance (LRTA) of each OTU (LRTA ≥ 30, LRTA ≥ 20, or LRTA < 20). Red text indicates q < 0.10. Points are colored by taxonomy represented heavily in top results, for p < 0.10. F., G., and S. in taxonomy labels indicate that the level of taxonomy is family, genus, or species