The scale shows time in coalescent units. The phylogeny with Seliciclib solubility dmso recombination correction also shows for each isolate its proportion of ancestry for each genetic cluster determined by the Structure analyses. For K = 2 and K = 6, the different colors represent each cluster. The proportion of color shading for each bar represents the proportion of ancestry for the respective cluster. Vertical bars show the isolates assigned to clusters A and B when K = 2. Asterisk refers to bovine isolates; # refers to feline isolate. The amount of recombination in bacteria can be quantified using two ratios: (i) the ratio of the frequency at which recombination occurs relative to mutation (ρ/θ),

and (ii) the ratio of the rates at which nucleotides become VDA chemical inhibitor substituted

as a result of recombination and mutation (r/m). The latter ratio accounts for length and nucleotide diversity of imported fragments and therefore contains more information regarding the evolutionary impact of recombination [69]. Using ClonalFrame, we calculated these ratios to be: ρ/θ = 0.1 and r/m = 1.5, with the latter ratio indicating that recombination exceeded point mutation. Vos and Didelot [70] calculated r/m for 48 diverse species of bacteria, and their results revealed a wide range of values (63.6 – 0.02). r/m for S. canis ranked 25th in this distribution (approximately AZD5582 in the middle). However, the average of the 48 rates was 7.7, suggesting a below average rate of recombination for S. canis when compared to these species of bacteria. When compared to the two Streptococcus species in the distribution, S. canis was much lower: S. pneumoniae = 23.1 (6th), S. pyogenes = 17.2 (8th). Similar results were obtained when ρ/θ for S. canis was compared to other Streptococcus species: S. uberis = 17.2 [71], S. pneumoniae = 23.1 [72]. We expanded the evolutionary analysis by also applying the parsimony-based approach e-BURST [73], which explores fine scale evolutionary relationships among STs. The ClonalFrame phylogeny and e-BURST results were generally concordant regarding the grouping of STs (Figure 3). The only

discrepancy was ST7, which showed an intermediate relationship between STs 9 and 10 in the phylogeny, ADAMTS5 but was not grouped within the same clonal complex (CC) as STs 9 and 10 (ST7 was not grouped into any of the four clonal complexes). Population structure was further examined using the Bayesian clustering approach implemented in the program Structure [74, 75]. The number of clusters K was estimated by calculating the ad hoc statistic ΔK, which is a measure of the second order rate of change of the probability of the data L(K) for each value of K[76] (see Methods for a full explanation of the approach). The analysis showed the optimum number of genetic clusters (K) to be two (A and B) (Figure 3 and Additional file 6). All four clonal complexes and ST8 were grouped into cluster A, whereas cluster B contained STs 6, 14, and 15.