Comparative Genomic Analysis Across Bacterial Pathogens

Analysis of the full set of PMEN1 sequences used the alignment from (26 (link)); 11 closely-related isolates were extracted as a subsample for comparison with the output of ClonalFrame. For the analysis of S. aureus ST239, 14 representatives from the South-East Asian clade were extracted from the larger alignment (49 (link)) for the equivalent comparative analysis. For the analysis of Helicobacter pylori, eight publically available complete genomes were selected from across the species that included both the most closely-related pair of isolates and the isolate most divergent from the rest of the sample, based on a previous analysis (50 (link)). These genomes were then aligned using progressiveMauve (51 (link)), generating a 1.8 Mb core genome alignment for analysis.
The resulting whole genome alignments were then analyzed using the default settings of Gubbins, except that the S. pneumoniae and S. aureus analyses were run until convergence. For S. pneumoniae and S. aureus, ClonalFrame (19 (link)) was also run using default settings, without estimating node ages, with a burn in chain length of 25 000 and a parameter estimation chain length of 25 000. For H. pylori, convergence was achieved when ClonalFrame was run without estimating node ages or theta, using a burn in chain length of 10 000 and a parameter estimation chain length of 10 000. Convergence was assessed through plotting the variation in parameter values over the course of the MCMC; these are shown in Supplementary Figures S4, S7 and S9.