Figure 3

Genome-wide phylogenetic analysis of 30 completely sequenced E. coli and Shigella strains. A, maximum likelihood phylogenetic tree, based on the concatenated nucleotide sequences of 341 orthologous CDSs from each of the strains (Additional file3: Dataset S1). The tree was constructed using RaxML with the JTT + GAMMA + Invariable sites model with 100,000 pseudoreplicates. The 341 CDS were selected from 345 CDS that were previously determined as being non-recombinogenic[22]. Four of 345 genes were removed because a subgroup of strains does not have these genes. B, maximum likelihood phylogenetic tree, constructed using genome-wide orthologous SNPs from 30 complete genomes. Pairwise comparisons of all genome sequences were carried out using NUCmer from the MUMer package[46] and highly similar regions (repeated sequences) were removed from the analysis. Orthologous SNPs are defined as SNPs present within the remaining alignments among all genomes. Only those SNPs present within the coding (CDS) regions were used for further phylogenetic analysis. The best substitution model (GTR + G) for the analysis was selected using ModelTest[47], and the tree was constructed using RAxML[48] with100,000 bootstrap replicates. The EHEC strains (EcO157: Sakai, Xuzhou 21, EDL933, EC4115 and TW14359; EcO145: RM13514 and RM13516; EcO26: 11368; EcO103: 12009; EcO111: 11128) and the STEC strain (EcO104: 2011C-3493) are in bold. Scale bar: number of substitutions per base. Only bootstrap value <= 95 are displayed.