Review A polyphasic strategy incorporating genomic data for the taxonomic description of novel bacterial species Dhamodharan Ramasamy, Ajay Kumar Mishra, Jean-Christophe Lagier, Roshan Padhmanabhan, Morgane Rossi, Erwin Sentausa, Didier Raoult and Pierre-Edouard Fournier Correspondence Pierre-Edouard Fournier pierre-edouard.fournier@ univ-amu.fr Unite ´ de Recherche sur les Maladies Infectieuses et Tropicales Emergentes URMITE, Institut Hospitalo-Universitaire Me ´ diterrane ´ e-Infection, Aix-Marseille Universite ´ , UMR63, CNRS 7278, IRD 198, INSERM U1095, Faculte ´ de Me ´ decine, 27 Bd Jean Moulin, 13005 Marseille, France Currently, bacterial taxonomy relies on a polyphasic approach based on the combination of phenotypic and genotypic characteristics. However, the current situation is paradoxical in that the genetic criteria that are used, including DNA–DNA hybridization, 16S rRNA gene sequence nucleotide similarity and phylogeny, and DNA G+C content, have significant limitations, but genome sequences that contain the whole genetic information of bacterial strains are not used for taxonomic purposes, despite the decreasing costs of sequencing and the increasing number of available genomes. Recently, we diversified bacterial culture conditions with the aim of isolating uncultivated bacteria. To classify the putative novel species that we cultivated, we used a polyphasic strategy that included phenotypic as well as genomic criteria (genome characteristics as well as genomic sequence similarity). Herein, we review the pros and cons of genome sequencing for taxonomy and propose that the incorporation of genome sequences in taxonomic studies has the advantage of using reliable and reproducible data. This strategy, which we name taxono-genomics, may contribute to the taxonomic classification of bacteria. Introduction Taxonomic information is essential, as it enables scientists to understand the biodiversity and relationships among living organisms from different ecosystems (Gevers et al., 2005). For prokaryotes, taxonomy plays an essential role in enabling the reliable identification of microbial strains from clinical or environmental specimens (Moore et al., 2010). Bacterial taxonomy was initiated in the late 19th century. Initially, bacteria were classified on the basis of basic phenotypic markers such as morphology, growth re- quirements or pathogenic potential (Lehmann & Neumann, 1896). Later, physiological and biochemical properties of bacteria were also used for this purpose (Orla-Jensen, 1909; Buchanan, 1955). Between the 1960s and the 1980s, chemotaxonomy (Minnikin et al., 1975), numerical tax- onomy and DNA–DNA hybridization techniques (Brenner et al., 1969; Johnson, 1991) were used. In the 1980s, the advent of DNA amplification and sequencing techniques, in particular of the 16S rRNA gene, constituted a major step forward by facilitating bacterial classification (Gu ¨rtler & Mayall, 2001; Coenye & Vandamme, 2004; Konstantinidis & Tiedje, 2007). Then, starting in the mid-1990s, whole- genome sequencing constituted a revolution by giving access to the complete genetic information of a strain (Janssen et al., 2003). Despite this tremendous progress and the various genome-based methods that were developed and proposed for taxonomic purposes, including multilocus sequence analysis and average nucleotide identity (ANI), whole- genome analysis (Stackebrandt et al., 2002; Rossello ´ -Mora, 2005; Goris et al., 2007; Konstantinidis & Tiedje, 2007; Yarza et al., 2008) has not as yet been accepted as a source of taxonomic information. Therefore, currently, routine identification of a bacterial strain is most often obtained by comparing its phenotypic and/or molecular characteristics with those of type strains of previously described species. In the case of an unusual strain, several minimal standards are used to determine whether it fulfils the requirements to be assigned to a novel taxon. However, the validity of this system is debated, as several of the criteria used were selected empirically, and some of the methods used are time- and money- consuming, are not accessible to all laboratories and lack intra- and inter-laboratory reproducibility (Stackebrandt & Ebers, 2006; Rossello ´ -Mora, 2006). Abbreviations: ANI, average nucleotide identity; DDH, DNA–DNA hybridization; HGT, horizontal gene transfer. International Journal of Systematic and Evolutionary Microbiology (2014), 64, 384–391 DOI 10.1099/ijs.0.057091-0 384 057091 G 2014 IUMS Printed in Great Britain