persica 16-FTT0376a, 17-FTT0523a, 20-ISFtu2b and 28-pdpDb.  Amplifies only F. tularensis (only when including the probe). 16-FTT0376a and 17-FTT0523a  Amplifies F. tularensis subsp. mediasiatica, F. tularensis subsp.holarctica and 6/7 F. tularensis subsp. novicida. 28-pdpDb  Amplifies isolates from all clade 1 species as well as W. persica. 20-ISFtu2b Marker with missing sequences as well as mismatches in almost all subspecies represented. 21-ISFtu2a Navitoclax manufacturer Successful amplification was defined as having a primer score below two in both the forward

and reverse primers. a Have associated TaqMan probe which is not considered here. bDetection by variable-length amplicon which is not considered here. cScore of F.noatunensis subsp orientalis <2. Evaluation of sample-sequencing approaches for phylogenetic analyses In the phylogenetic comparison analysis, we focused not only on the entire Francisella genus, but also selleck chemicals llc analysed clades 1 and 2 separately. These sub-populations exhibit different lifestyles and environmental niches and are therefore of interest to different scientific fields [3, 7, 18]. The differences between the poorest and best resolved single marker topologies of the entire genus compared to the whole-genome reference topology (Figure 2) are highlighted in Figure 3A-C. All topologies are shown in Additional File 2. The parameter estimates of the phylogenetic

analysis are summarised in Additional File 3. In general for the analysis of the entire genus, the optimal substitution model was parameter rich, i.e. typically the generalised Thiamine-diphosphate kinase time-reversible (GTR) [31] or Hasegawa-Kishino-Yano (HKY85) [32] models with either invariant sites parameter (α) or rate heterogeneity over sites (Г). Moderate or even low parameter-rich substitution models were favoured in the separate clade analyses, in particular for clade 1, where Jukes-Cantor (JC) [33] or HKY85 models were found to be the optimal choice without α or Г. For clade 2, it was important to include the proportion of invariant sites parameter in the analyses, because of detected recombination events [3].

Figure 2 Whole-genome SNP phylogeny. The whole-genome phylogeny for 37 Francisella strains obtained with model averaging Alpelisib mouse implemented in jModelTest using PhyML software. The removed part of the branches connecting clade 1 and 2 covers a genetic distance of 0.03. Figure 3 Single-marker phylogenies. Single-marker phylogeny of the Francisella genus: (A) highest ranked marker 08-fabH, (B) lowest ranked marker 33-rpoB, and (C) whole-genome phylogeny. Rank is based on difference in resolution between alternative and whole-genome topology. Throughout the study, to facilitate the phylogeny comparisons, we made use of two metrics: degree of incongruence (inc) and difference in resolution (res). The two topologies compared were the reference topology, obtained from whole genome data, and the single-sequence or the concatenated marker sequences topology.