Extended Data Fig. 3: Overview of sputum metagenomic sequencing.

a) Nonpareil curves of all samples showing the estimated average coverage as a function of sequencing effort. For each curve, the hollow circle represents the actual sequencing effort for that sample and its corresponding estimated average coverage. Samples with estimated microbial coverage below 70% are colored in blue. b) Barplot showing the correlation coefficient between the shared genera identified from 16S rRNA gene-based amplicon data (relative abundance > =1e-4) and metagenomic data. Significant correlations were found in 45 out of 51 genera (P < 0.001, Pearson correlation). c) A phylogenetic diagram showing the bacterial genera identified in amplicon or metagenomic data. Shown from the outside inward are the presence or absence of the genus in metagenomic data, in 16S rRNA gene-based amplicon data, and its corresponding phylum. A greater number of genera were identified in amplicon than metagenomic data. d) Histogram showing the distribution on the proportion of viral reads of the overall microbial community for all 1,128 samples. e) Piechart showing the top 10 most abundant viral species in the observed viral community across all 1,128 samples. f) Rarefaction plot showing the number of viral species detected with the increment of sequencing depth. The exact P-values for comparisons in b) are provided in Supplementary Table 18. *** P < 0.001, ** P < 0.01, * P < 0.05, + P < 0.1.