PHYLOGENETIC INFERENCE IN BACTERIA: PHYLOGENOMICS VS MULTILOCUS SEQUENCE TYPING (MLST) SCHEMES

NOELIA FLORIDIA-YAPUR; FANNY RUSMAN; PAULA G. RAGONE; PATRICIO DIOSQUE; LEONARDO ACUÑA; NICOLÁS TOMASINI

Salta

Congreso; The LV Annual SAIB Meeting and XIV PABMB Congress; 2019

Sociedad Argentina de Bioquímica y Biología Molecular

The whole-genome sequence era brought a massive amount of information that helped to elucidate or clarify phylogeny relations of severalorganisms. It has been proposed that genome-based phylogeny is more accurate than phylogeny based in a few number of genes as MLST-basedphylogeny. However, much is not always better. Here we studied the intraspecific phylogeny of 9 bacteria species, with more than 30 completegenomes sequenced: Burkholderia pseudomallei, Campylobacter jejuni, Chlamydia trachomatis, Helicobacter pylori, Klebsiella pneumoniae,Listeria monocytogenes, Salmonella enterica, Staphylococcus aureus and Streptococcus pyogenes. The phylogeny was inferred using, on the onehand, the complete genome sequences of different strains of these species and, on the other hand, their MLST schemes. Complete genomes/MLSTgenes were aligned. Maximum likelihood trees were constructed with FastTree2 for both complete genomes and concatenated MLST loci. Inaddition, genome alignments were fragmented every 5000 nt and maximum likelihood trees were built for every genome fragment (individualtrees). The phylogenetic incongruence between the complete genome or MLST concatenated sequences tree and the individual trees was analyzedusing MLSTest. In genome trees, most branches showed a high branch support; however, a high number of branches also showed a high percentageof topologically incongruent individual trees. For example, 90% (37/41) of the branches of the genome tree of B. pseudomallei showed a high branch support (more than 0.8); 31 of those branches had the maximum branch support (1); however, 51% of such branches are probably artifactsbecause most genome fragments are topologically incompatible with them. Interestingly, genome and MLST trees showed similar levels ofincongruence in the phylogeny of each bacteria specie. Both genome and MLST approaches showed that C. trachomatis and S. aureus have a treelike evolutionary history (low levels of internal incongruence). Instead B. pseudomallei and S. pyogenes show high levels of incongruence probablycaused by horizontal gene transfer. Consequently, their evolutionary history is better represented by a network. Concluding, the high branchsupport obtained in genome phylogenies could be an artifact probably caused by the data size. In addition, our analyses showed that MLSTphylogeny is a valid method to address intraspecific phylogeny and that some species phylogenies are not properly represented by trees.