Principal Coordinates Analysis of unweighted UniFrac distances for 16S rDNA V1–V3 sequence data.

Beta diversity patterns were explored using Principal Coordinates Analysis (PCoA). Samples were rarefied to 6,596 reads/sample for Figs. 2A and 2C and rarefied to 427 reads/sample for 2B due to the smaller number of available reads per U.S. adult. The near full-length 16S rDNA data were not included in this analysis. (A) The sample symbols for Bangladeshi and U.S. children are colored by country. Only the data from the children (with the exception of BK8F and BK9F) were included in this PCoA plot. (B) Nine U.S. adults were included in this analysis (one time point per adult). The fecal communities for all three subjects (Subjects A, B, and C) in the Eckburg et al. study (near full-length 16S rDNA sequences) and all 6 subjects (Subjects A, B, C, D, E, F) in the Dethlefsen et al. studies (V1–V3 16S rDNA sequences) were included in the analysis [11], [25], [26]. The near full-length 16S rDNA sequences were filtered in silico to the V1–V3 regions using ARB. Data from all of the children were included in this analysis. (C) Same as (A) but with each sample symbol colored-coded by subject.