Whole-genome bacterial classification : whole genome taxonomic reclassification and a universal prokaryotic identifier scheme

Binomial taxonomic nomenclature is problematic for prokaryotes. Much existing classification has origins in polyphasic and phenotypic classifications that do not reflect relatedness at molecular level, and is under continual active revision that results in widespread confusion in literature, databases, and historical collections. Taxonomic classification nevertheless remains central to many areas of significant public impact, including development of political policy for legislation, and border control that aims to reduce disease risks to agriculture from pathogenic bacteria. To meet policy goals effectively with diagnostic tools and associate disease risk with identity, historical classifications need to be revisited. We use measures of genomic relatedness and a graph decomposition approach to subdivide enterobacterial plant pathogens into groupings (cliques) based only on inherent properties of their complete genomes. These groups require no arbitrary thresholding and are stable to introduction of new sequences. They are capable of providing a basis for universal indexing of prokaryotes and probabilistic estimates of risk conditioned on clique membership