Skip to main content
. 2023 Mar 2;14:1191. doi: 10.1038/s41467-023-36633-7

Fig. 2. Correlation between median diversity and abundance of ARGs in healthy adult gut metagenomes in a country and antibiotic consumption rates.

Fig. 2

a Median total gut ARG abundance (cpg) in healthy antibiotic free adults for countries with > 10 samples. b, c Correlations between median total ARG abundance and the per capita rate of antibiotic consumption (DDD per 1000), using CDDEP consumption statistics in (b) and WHO in (c). Pearson’s correlation tests gave r = 0.89, p = 0.00023 for CDDEP (b) and r = 0.65, p = 0.040 for WHO (c), excluding China from the CDDEP correlations. d Median rarefied (100 ± 10 Gbp subsamples) ARG_cluster95 richness across countries. e, f Correlations between ARG_cluster95 richness per 100 Gbp and the per capita rate of antibiotic consumption for CDDEP (e) and WHO (f) estimates. Pearson’s correlation tests r = 0.57, p = 0.11 for CDDEP (e) and r = 0.85, p = 0.0073 for WHO (f), excluding China from the CDDEP correlations. Vertical lines range from the 25th percentile to 75th percentile (b, c, e, f). Country-level ARG statistics are given in Table S3. Correlation tests in Table S4. Novelty of the ARG clusters with respect to RefSeq is visualized in Fig. S1. Rarefaction curves of ARG cluster richness in human microbiomes are shown in Fig. S2. . Linear trend line was determined by a generalized linear model using ggplot2 R package, the shaded area represents 95% confidence interval. Country name abbreviations and metagenome sample number: AUT (n = 16) - Austria, CAN (n = 35) - Canada, CHN (n = 209) - China, DEU (n = 103) - Germany, DNK (n = 230) - Denmark, ESP (n = 139) - Spain, FRA (n = 62)- France, ISR (n = 937) - Israel, ITA (n = 33) - Italy, KAZ (n = 168) - Kazakhstan, MDG (n = 112) - Madagascar, NLD (n = 468) - Netherlands, SWE (n = 109) - Sweden, USA (n = 115) - United States. The number of rarefactions performed to derive each box plot and range bar shown in d–f: n = 99. In the box plots shown in a and d, the box spans from 25th to 75th percentiles, the line inside the box is the median, and the whisker spans from the minimum to the maximum values.