Abstract
The genus Pseudomonas currently contains 144 species, making it the genus of Gram-negative bacteria that contains the largest number of species. Currently, multilocus sequence analysis (MLSA) is the preferred method for establishing the phylogeny between species and genera. Four partial gene sequences of housekeeping genes (16S rRNA, gyrB, rpoB, and rpoD) were obtained from 112 complete or draft genomes of strains related to the genus Pseudomonas that were available in databases. These genes were analyzed together with the corresponding sequences of 133 Pseudomonas type strains of validly published species to assess their correct phylogenetic assignations. We confirmed that 30% of the sequenced genomes of non-type strains were not correctly assigned at the species level in the accepted taxonomy of the genus and that 20% of the strains were not identified at the species level. Most of these strains had been isolated and classified several years ago, and their taxonomic status has not been updated by modern techniques. MLSA was also compared with indices based on the analysis of whole-genome sequences that have been proposed for species delineation, such as tetranucleotide usage patterns (TETRA), average nucleotide identity (ANIm, based on MUMmer and ANIb, based on BLAST) and genome-to-genome distance (GGDC). TETRA was useful for discriminating Pseudomonas from other genera, whereas ANIb and GGDC clearly separated strains of different species. ANIb showed the strongest correlation with MLSA. The correct species classification is a prerequisite for most diversity and evolutionary studies. This work highlights the necessity for complete genomic sequences of type strains to build a phylogenomic taxonomy and that all new genome sequences submitted to databases should be correctly assigned to species to avoid taxonomic inconsistencies.