Extended Data Fig. 7: Relationships between similarity in strains’ interaction profiles and their phylogenetic distance.
From: A host–microbiota interactome reveals extensive transkingdom connectivity

a, We computed a phylogenetic tree over 108 genomes of tested strains based on ~ 400 broadly distributed protein families. We compared distances in this tree with similarity of strains’ interaction profiles using Spearman correlation (n = 5,565 strain pairs). Phylogenetic distance is expressed in units of amino acid substitutions per amino acid site. Interaction similarity was measured as the Jaccard overlap score between strains’ sets of human protein binding partners (ignoring strains with no binding partners). b, We separately considered the subset of n = 907 strain pairs with phylogenetic distance <0.02 substitutions per site, which was largely synonymous with a conspecific relationship in taxonomy. In both regimes, interaction similarity and phylogenetic distance were strongly and significantly negatively correlated. In both cases a two-tailed Mantel test with 104 permutations with FDR adjustments was performed.