Assessment of agreement between the in vivo results from Wu et al. [48] and gut-specific (“bodysite”) vs. gut prevalence (“overall”) phylogenetic models.

The background sets for enrichment tests were defined as follows: “all tested” (all gene families for which a phylogenetic model was fit), “Bacteroides (core or variable)” (all gene families with at least one representative in Bacteroides genome cluster pangenomes), “Bacteroides (core only)” (gene families that were present in all Bacteroides genome cluster pangenomes), “Bacteroides (variable only)” (gene families present in some but not all Bacteroides genomes clusters), and “Bacteroides thetaiotaomicron only” (only gene families present in Bacteroides thetaiotaomicron). The “all tested” background sets are further classified as “all tested (overall)” and “all tested (body site)” since the set of genes tested differed slightly. The p-values are from Fisher’s exact tests. These comparisons have been excerpted from the full set, which can be seen in S3 Table; q- values were calculated based on this full set of tests using the Benjamini-Hochberg method [89].