Fig. 3

Significant differences in centered log-ratio (CLR) relative abundance of prevalent taxa between geographic regions and ages across healthy control samples. Heatmaps show significant taxonomic associations with geographic location and age in healthy controls. In both, mean CLR-transformed relative abundance is shown via color encoding, with red indicating higher CLR abundance and blue indicating lower CLR abundance. A Taxa displaying significant associations with geographic location in healthy controls are shown in each column (N = 2387). Each row represents one study, with the URT sampling site annotated (NP = nasopharynx, OP = oropharynx). Geographic region per study is shown via the color bar to the left of the heatmap. Significance was determined by multiple regression, correcting for URT sampling site, sequencing method, and 16S hypervariable region, with FDR-corrected two-tailed p-value < 0.05. B Taxa significantly associated with age are shown, for samples with available metadata for age (N = 554). Significance was determined by ANCOVA, treating age as a continuous variable, correcting for geographic region, URT sampling site, sequencing method, and 16S hypervariable region, with FDR-corrected two-tailed p-value < 0.05