Two isolates (H063920004 and H091960009) were sequenced with different technologies. H063920004 with Illumina paired end, Illumina mate-paired and Roche 454, H091960009 with Illumina paired end and Illumina mate-paired. There were

no SNP differences between the sequences of these replicate samples demonstrating that the protocol used for calling SNP variants is both robust and consistent. There were three isolates of ST47 (labelled ST47, LP-617 and Lorraine), two from the UK and one from France, each isolated in a different year between 2003 and 2006. These differed by just four SNPs. Two ST42 isolates, from the UK and USA (labelled #Barasertib in vivo randurls[1|1|,|CHEM1|]# ST42 and Wadsworth), were isolated 20 years apart and only exhibited 20 SNP differences. In contrast two ST1 isolates, a representative of the ‘Paris’ strain and a UK strain sequenced as part of another study, were isolated within 2 years of each other yet these exhibited 280 SNP differences. These results show that lineages of L. pneumophila contain differing levels of observable diversity. There are several evolutionary scenarios that could be postulated ITF2357 in vitro as explanations for these observed differences. A lineage that occupies a niche where there is strong purifying selection will be less diverse. Conversely a lineage that is the result of rapid expansion

within a previously unoccupied niche will tend to be more diverse. One likely scenario is that ST1 is a successful clonal lineage that emerged before the ST47 lineage and therefore has had more time to diversify by genetic drift. It is also possible that each lineage of L. pneumophila will be subject to differing selection pressures when infecting a human host, even though this is effectively an evolutionary dead-end. One possible scenario is that the majority of ST1 strains

and a limited number of sub-lineages of ST47 cause disease in humans. If this is the case then a likely explanation is that the common ancestor of the ST1 lineage was able to infect the human species and the ancestor of the ST47 lineage did not replicate effectively in a human host. Subsequently a minority of descendents of the ST47 lineage have acquired PIK3C2G the ability (through mutation, gene loss or acquisition) to cause human infection. Differentiating between these putative evolutionary scenarios will be difficult and will require a greater understanding of the effects of diversity within the lineages of L. pneumophila sampled from the environment and human infections. When examining the output from the Splits Tree analysis, the more splits observed, the more recombination or HGT is likely to have taken place. The majority of clades in the tree show a branching network structure suggestive of frequent recombination. The Phi test for recombination as implemented in SplitsTree also showed evidence for recombination (p = 0.0). The exceptions are the clade(s) containing ST136/154 and ST707.