Fig. 7: Eight clinically important pathogens and their R-MIC dynamics.
From: Seeking patterns of antibiotic resistance in ATLAS, an open, raw MIC database with patient metadata
A Each panel shows the (R, dR/dt) phase plane (x-axis: 2017 data, the most recent value in ATLAS; the normalised \({\log }_{2}({\mathsf{MIC}})=0\) axis represents the clinical breakpoint, see “Methods”). Data are means and crosshairs are s.d. under synthetic ATLAS replication with n = 50 (“Methods”). Acinetobacter baumannii exhibits clinical resistance to all ATLAS antibiotics and the other phase planes visibly skew towards resistance. Labels C0-C8 refer to clinical observations made in the main text. Data with low year-year correlations satisfying τ < 1/4 (Fig. 3 and Supplementary Fig. 10) are circled red. B A 2d heat map counts which regions contain 544 pathogen-antibiotic pairs in the (R, dR/dt) phase plane. C Example changes in R-MIC using global data for Pseudomonas aeruginosa and 3 carbapenems: doripenem MICs are increasing from sub- to super-breakpoint whereas meropenem appears stable. Linear regression of R (red dashes) indicate the R-MIC of imipenem is slowly decreasing whereas, in fact, inspection of the underlying data shows this R-cluster is merging with an S sub-population as both achieve super-breakpoint dosages and the population transitions from a bimodal to unimodal MIC distribution.
