Fig. 1.
From: Large-scale genomic analysis of antimicrobial resistance in the zoonotic pathogen Streptococcus suis
Candidate AMR determinants explain most of the variation in MIC. Histograms of log transformed MIC measures for each of our 16 different antibiotics, across our panel of 678 S. suis isolates. Antibiotics are coloured by their class (Beta: beta-lactams; MLSB: macrolide-lincosamide-streptogramin B; Tetra: tetracyclines; Fluoro: fluoroquinolones; Amino: aminoglycoside; Pleuro: pleuromutilin; TMP: trimethoprim and Phen: phenicol). In the 16 square panels, the left-hand histograms (labelled 0) show the MIC values for isolates that carry no determinant for that antibiotic class, while the right-hand histograms (labelled 1+) show the MIC values for isolates carrying one or more such determinant. If all resistance determinants perfectly explain MIC, then we expect to see histogram distributions on the bottom left (low MIC, no determinant) and the top right (high MIC, presence of candidate determinant(s)). For the first antimicrobial in each class, we show the number of candidate AMR determinants in square brackets, along with the number of isolates where candidate determinants are absent or present. r2 values show the proportion of the variance explained in a standard ANOVA by the presence of one or more candidate determinant