Resolving phylogenetic relationships for streptococcus mitis and streptococcus oralis through core- A nd pan-genome analyses

Abstract

Taxonomic and phylogenetic relationships of Streptococcusmitis and Streptococcus oralis have been difficult to establish biochemically and genetically. We used core-genome analyses of S. mitis and S. oralis, as well as the closely related species Streptococcus pneumoniaeandStreptococcusparasanguinis, toclarify thephylogenetic relationshipsbetweenS.mitisandS.oralis,aswell aswithin subcladesof S.oralis.All S. mitis (n=67), S.oralis (n=89), S.parasanguinis (n=27), and 27S .pneumoniaegenomeassemblieswere downloaded fromNCBI and reannotated.Allgenesweredelineatedinto homologous clusters andmaximum-likelihood phylogenies built from putatively nonrecombinant core gene sets. Population structurewas determined using Bayesian genome clustering, and patristic distance was calculated between populations. Population-specific gene content was assessed using a phylogenetic-based genome-wide association approach. Streptococcusmitis and S. oralis formed distinct clades, but speciesmixing suggests taxonomic misassignment. Patristic distance between populations suggests that S. oralis subsp. dentisani is a distinct species, whereas S. oralis subsp. tigurinus and subsp. oralis are supported as subspecies, and that S.mitis comprises two subspecies. None of the geneswithin the pan-genomes of S.mitis and S. oralis could be statistically correlatedwith either, and the dispensable genomes showed extensive variationamongisolates. These are likely important factors contributingtoestablishedoverlapin biochemical characteristics for these taxa. Based on core-genome analysis, the substructure of S. oralis and S.mitis should be redefined, and species assignmentswithin S. oralis and S. mitis should be made based on whole-genome analysis to be robust to misassignment.