Fig. 3. Substrain variation and plasmid sharing in Enterobacteriaceae species.
a,b, Line charts showing the allele frequencies of metagenomics-derived SNVs across timepoints for E. coli (a) and K. pneumoniae (b), for subject 1674-T. SNVs that belong to different substrain clusters are coloured accordingly, and a potential model for the corresponding haplotypes is shown in Supplementary Fig. 7. All genome-wide SNVs detected at all the depicted timepoints are shown here. The blue box indicates the CPE-negative timepoint. c,d, Corresponding genomic distribution for SNVs belonging to cluster 1 or cluster 2 and those that are fixed at all timepoints (‘Shared’) or just at the CPE-negative timepoint (‘V05 unique’). The associated fraction describes the number of SNVs in a grouping that are common with CPE isolates, in relation to the total number of SNVs in this grouping. e,f, Relative abundance of E. coli and K. pneumoniae in the gut metagenome for subject 1674-T across timepoints. g, Visualization of amino acid changes (red) seen in a key polysaccharide utilization protein (lacZ, pdb structure 4DUX chain A, identified from Table 1 and Extended Data Table 2) based on metagenomics-derived SNVs that changed in frequency over time and across multiple individuals. SNVs are found primarily on the protein surface, with six SNVs located on the activating interface responsible for tetramerization (teal). h, Diagram depicting plasmid sharing between E. coli and K. pneumoniae strains at various timepoints for subject 1674-T. Plasmid sequences were clustered at 95% similarity (Supplementary Fig. 8), and a representative plasmid for each cluster is shown in the Venn diagram as an approximately sized circle, with plasmid name and size.