The comparisons varied in inc, and sometimes considerably so. In the analysis of the entire genus, the 37-trpE topology did not HDAC inhibitor exhibit any incongruence compared to the reference (inc = 0), although the resolution was poor. For other markers, such as 08-fabH, 27-parC, 03-16 s + ItS + 23 s, 04-16 s + ItS + 23 s, 25-mutS and 36-tpiA, the topology comparisons indicated few mismatched bipartitions (inc < 0.25), whereas the opposite result was found for 11-fopA-in, 29-pgm and 30-prfB (inc > 0.35). As expected, for some single-marker topologies, particularly those with the lowest inc scores, the SH test did not HSP990 solubility dmso reject congruence compared to the reference phylogeny. Separate clade 1 topologies exhibited
a lower average incongruence than topologies of the entire genus (incclade1 = 0.139 vs. incgenus = 0.258, p = 6.6e-05) and clade 2 topologies (incclade1 = 0.139 vs. incclade2 = 0.238, p = 0.0149). In several cases, clade 1 topologies were totally congruent with no mismatched bipartitions. Some of these topologies were also congruent in clade 2: (01-16S,
03-16 s + ItS + 23 s, 04-16 s + ItS + 23 s, 07-dnaA, 08-fabH, 22-lpnA, 24-lpnB, 25-mdh, 27-parC, 30-prfB, 31-putA, 35-tpiA, 36-tpiA, 37-trpE and 38-uup). The low level of incongruence was verified by the results of the SH-test, which showed that congruence in the topology comparisons could not be rejected with the exception of 19-iglC. Reported incongruences in clade 1 mostly occurred in F. novicida. Most assignments deviating from the reference in clade 2 were due to misplacements learn more of subspecies F.
philomiragia and F. noatunensis subsp. noatunensis. In the separate analysis of clade 1, most strains not assigned according to the reference were due to poor resolution, notably topologies of markers 32-rpoA, 37-trpE, 25-mdh, 24-lpnB and 19-iglC. The average resolution (res) in topologies of clade 1 was significantly higher than clade 2 (resclade1 = 0.723 vs. resclade2 = 0.604, p = 0.003) and the entire genus (resclade1 = 0.723 vs. resgenus = 0.664, p = 0.010). The correlations between the incongruence and resolution Tenoxicam metrics were ρ = 0.405 and ρ = 0.484 for clade 1 and 2, respectively. Figure 4 shows the difference in comparison metrics and average bootstrap support (boot) when markers were randomly concatenated and an optimised combination of markers was selected. Table 4 lists optimal sets of two to seven markers for use in studies of the Francisella genus. Summary statistics of the optimal combinations of markers in the entire genus are summarised in Additional File 5. Results of the optimisation analyses of the separate clades are not shown. Compared to random concatenation of sequence markers, the Francisella genus topology from an optimised set of markers reduced the difference in resolution by on average 50 – 59% and totally eliminated incongruences.