High-Quality Whole-Genome Sequences for 77 Shiga Toxin-Producing Escherichia coli Strains Generated with PacBio Sequencing

A prophage encoded ribosomal RNA methyltransferase regulates the virulence of Shiga-toxin-producing Escherichia coli (STEC)

Shiga toxin (Stx) released by Shiga toxin producing Escherichia coli (STEC) causes life-threatening illness. Its production and release require induction of Stx-encoding prophage resident within the STEC genome. We identified two different STEC strains, PA2 and PA8, bearing Stx-encoding prophage whose sequences primarily differ by the position of an IS629 insertion element, yet differ in their abilities to kill eukaryotic cells and whose prophages differ in their spontaneous induction frequencies. The IS629 element in ϕPA2, disrupts an ORF predicted to encode a DNA adenine methyltransferase, whereas in ϕPA8, this element lies in an intergenic region. Introducing a plasmid expressing the methyltransferase gene product into ϕPA2 bearing-strains increases both the prophage spontaneous induction frequency and virulence to those exhibited by ϕPA8 bearing-strains. However, a plasmid bearing mutations predicted to disrupt the putative active site of the methyltransferase does not complement either of these defects. When complexed with a second protein, the methyltransferase holoenzyme preferentially uses 16S rRNA as a substrate. The second subunit is responsible for directing the preferential methylation of rRNA. Together these findings reveal a previously unrecognized role for rRNA methylation in regulating induction of Stx-encoding prophage.

HycDemux: a hybrid unsupervised approach for accurate barcoded sample demultiplexing in nanopore sequencing

DNA barcodes enable Oxford Nanopore sequencing to sequence multiple barcoded DNA samples on a single flow cell. DNA sequences with the same barcode need to be grouped together through demultiplexing. As the number of samples increases, accurate demultiplexing becomes difficult. We introduce HycDemux, which incorporates a GPU-parallelized hybrid clustering algorithm that uses nanopore signals and DNA sequences for accurate data clustering, alongside a voting-based module to finalize the demultiplexing results. Comprehensive experiments demonstrate that our approach outperforms unsupervised tools in short sequence fragment clustering and performs more robustly than current state-of-the-art demultiplexing tools for complex multi-sample sequencing data.

Comparative genomic and phenotypic analyses of the virulence potential in Shiga toxin-producing Escherichia coli O121:H7 and O121:H10

Shiga toxin-producing Escherichia coli (STEC) O121 is among the top six non-O157 serogroups that are most frequently associated with severe disease in humans. While O121:H19 is predominant, other O121 serotypes have been frequently isolated from environmental samples, but their virulence repertoire is poorly characterized. Here, we sequenced the complete genomes of two animal isolates belonging to O121:H7 and O121:H10 and performed comparative genomic analysis with O121:H19 to assess their virulence potential. Both O121:H7 and O121:H10 strains carry a genome comparable in size with the O121:H19 genomes and belong to phylogroup B1. However, both strains appear to have evolved from a different lineage than the O121:H19 strains according to the core genes-based phylogeny and Multi Locus Sequence Typing. A systematic search of over 300 E. coli virulence genes listed in the Virulence Factor DataBase revealed a total of 73 and 71 in O121:H7 and O121:H10 strains, respectively, in comparison with an average of 135 in the O121:H19 strains. This variation in the virulence genes repertoire was mainly attributed to the reduction in the number of genes related to the Type III Secretion System in the O121:H7 and O121:H10 strains. Compared to the O121:H19 strains, the O121:H7 strain carries more adherence and toxin genes while the O121:H10 strain carries more genes related to the Type VI Secretion System. Although both O121:H7 and O121:H10 strains carry the large virulence plasmid pEHEC, they do not harbor all pEHEC virulence genes in O121:H19. Furthermore, unlike the O121:H19 strains, neither the O121:H7 nor O121:H10 strain carried the Locus of Enterocyte Effacement, OI-122, nor the tellurite resistance island. Although an incomplete Locus of Adhesion and Autoaggregation (LAA) was identified in the O121:H7 and O121:H10 strains, a limited number of virulence genes were present. Consistently, both O121:H7 and O121:H10 strains displayed significant reduced cytotoxicity than either the O157:H7 strain EDL933 or the O121:H19 strain RM8352. In fact, the O121:H7 strain RM8082 appeared to cause minimal cytotoxicity to Vero cells. Our study demonstrated distinct evolutionary lineages among the strains of serotypes O121:H19, O121:H10, and O121:H7 and suggested reduced virulence potentials in STEC strains of O121:H10 and O121:H7.

Prevalence and Implications of Shiga Toxin-Producing E. coli in Farm and Wild Ruminants

Shiga-toxin-producing Escherichia coli (STEC) is a food-borne pathogen that causes human gastrointestinal infections across the globe, leading to kidney failure or even death in severe cases. E. coli are commensal members of humans and animals' (cattle, bison, and pigs) guts, however, may acquire Shiga-toxin-encoded phages. This acquisition or colonization by STEC may lead to dysbiosis in the intestinal microbial community of the host. Wildlife and livestock animals can be asymptomatically colonized by STEC, leading to pathogen shedding and transmission. Furthermore, there has been a steady uptick in new STEC variants representing various serotypes. These, along with hybrids of other pathogenic E. coli (UPEC and ExPEC), are of serious concern, especially when they possess enhanced antimicrobial resistance, biofilm formation, etc. Recent studies have reported these in the livestock and food industry with minimal focus on wildlife. Disturbed natural habitats and changing climates are increasingly creating wildlife reservoirs of these pathogens, leading to a rise in zoonotic infections. Therefore, this review comprehensively surveyed studies on STEC prevalence in livestock and wildlife hosts. We further present important microbial and environmental factors contributing to STEC spread as well as infections. Finally, we delve into potential strategies for limiting STEC shedding and transmission.

Inflammatory bowel disease-associated adherent-invasive Escherichia coli have elevated host-defense peptide resistance

Adherent-invasive Escherichia coli (AIEC) are isolated from inflammatory bowel disease (IBD) patients at a higher rate than from control patients. Using a collection of E. coli strains collected from Crohn's disease (CD), ulcerative colitis (UC), or non-IBD control patients, antibiotic and resistance to the antimicrobial peptides HBD-3 and LL-37 was assessed. Carriage of bacterial-encoded omptin protease genes was assessed by PCR and omptin protease activity was measured using a whole-cell based fluorescence assay. Elevated resistance to antibiotics and host defense peptides in IBD-associated AIEC were observed. IBD-associated strains showed increased (but statistically non-significant) antibiotic resistance. CD-associated strains showed greater (but statistically non-significant) resistance to HBD3-mediated killing while UC-associated strains showed statistically greater resistance to LL-37 mediated killing. High-level resistance to LL-37 was associated with carriage of omptin protease genes and with increased omptin protease activity. Antimicrobial host defense peptide resistance may be an adaptive feature of AIEC leading to enhanced pathogenesis during the initiation or progression of IBD.

Comparative Genomics Applied to Systematically Assess Pathogenicity Potential in Shiga Toxin-Producing Escherichia coli O145:H28

Shiga toxin-producing Escherichia coli (STEC) O145:H28 can cause severe disease in humans and is a predominant serotype in STEC O145 environmental isolates. Here, comparative genomics was applied to a set of clinical and environmental strains to systematically evaluate the pathogenicity potential in environmental strains. While the core genes-based tree separated all O145:H28 strains from the non O145:H28 reference strains, it failed to segregate environmental strains from the clinical. In contrast, the accessory genes-based tree placed all clinical strains in the same clade regardless of their genotypes or serotypes, apart from the environmental strains. Loss-of-function mutations were common in the virulence genes examined, with a high frequency in genes related to adherence, autotransporters, and the type three secretion system. Distinct differences in pathogenicity islands LEE, OI-122, and OI-57, the acid fitness island, and the tellurite resistance island were detected between the O145:H28 and reference strains. A great amount of genetic variation was detected in O145:H28, which was mainly attributed to deletions, insertions, and gene acquisition at several chromosomal “hot spots”. Our study demonstrated a distinct virulence gene repertoire among the STEC O145:H28 strains originating from the same geographical region and revealed unforeseen contributions of loss-of-function mutations to virulence evolution and genetic diversification in STEC.

Pathogenomes and variations in Shiga toxin production among geographically distinct clones of Escherichia coli O113:H21

Infections with globally disseminated Shiga toxin-producing Escherichia coli (STEC) of the O113:H21 serotype can progress to severe clinical complications, such as hemolytic uremic syndrome (HUS). Two phylogeographically distinct clonal complexes have been established by multi locus sequence typing (MLST). Infections with ST-820 isolates circulating exclusively in Australia have caused severe human disease, such as HUS. Conversely, ST-223 isolates prevalent in the US and outside Australia seem to rarely cause severe human disease but are frequent contaminants. Following a genomic epidemiology approach, we wanted to gain insights into the underlying cause for this disparity. We examined the plasticity in the genome make-up and Shiga toxin production in a collection of 20 ST-820 and ST-223 strains isolated from produce, the bovine reservoir, and clinical cases. STEC are notorious for assembly into fragmented draft sequences when using short-read sequencing technologies due to the extensive and partly homologous phage complement. The application of long-read technology (LRT) sequencing yielded closed reference chromosomes and plasmids for two representative ST-820 and ST-223 strains. The established high-resolution framework, based on whole genome alignments, single nucleotide polymorphism (SNP)-typing and MLST, includes the chromosomes and plasmids of other publicly available O113:H21 sequences and allowed us to refine the phylogeographical boundaries of ST-820 and ST-223 complex isolates and to further identify a historic non-shigatoxigenic strain from Mexico as a quasi-intermediate. Plasmid comparison revealed strong correlations between the strains' featured pO113 plasmid genotypes and chromosomally inferred ST, which suggests coevolution of the chromosome and virulence plasmids. Our pathogenicity assessment revealed statistically significant differences in the Stx_2a-production capabilities of ST-820 as compared to ST-223 strains under RecA-induced Stx phage mobilization, a condition that mimics Stx-phage induction. These observations suggest that ST-820 strains may confer an increased pathogenic potential in line with the strain-associated epidemiological metadata. Still, some of the tested ST-223 cultures sourced from contaminated produce or the bovine reservoir also produced Stx at levels comparable to those of ST-820 isolates, which calls for awareness and for continued surveillance of this lineage.

Characterisation of atypical Shiga toxin gene sequences and description of Stx2j, a new subtype

Shiga toxin (Stx) is the definitive virulence factor of Shiga toxin-producing Escherichia coli (STEC). Stx variants are currently organised into a taxonomic system of three Stx1 (a,c,d) and seven Stx2 (a,b,c,d,e,f,g) subtypes. In this study, seven STEC isolates from food and clinical samples possessing stx2 sequences that do not fit current Shiga toxin taxonomy were identified. Genome assemblies of the STEC strains was created from Oxford Nanopore and Illumina sequence data. The presence of atypical stx2 sequences were confirmed by Sanger sequencing, as were Stx2 expression and cytotoxicity. A strain of O157:H7 was found to possess stx1a and a truncated stx2a, which were originally misidentified as an atypical stx2. Two strains possessed unreported variants of Stx2a (O8:H28) and Stx2b (O146:H21). In four of the strains we found three Stx-subtypes that are not included in the current taxonomy. Stx2h (O170:H18) was identified in a Canadian sprout isolate; this subtype has only previously been reported in STEC from Tibetan Marmots. Stx2o (O85:H1) was identified in a clinical isolate. Finally, Stx2j (O158:H23 and O33:H14) was found in lettuce and clinical isolates. The results of this study expands the number of known Stx subtypes, the range of STEC serotypes, and isolation sources in which they may be found. The presence of the Stx2j and Stx2o in clinical isolates of STEC indicates that strains carrying these variants are potential human pathogens. Highlights Atypical Shiga toxin (stx) genes in Escherichia coli were sequenced. Two new variants of stx2a and stx2b are described. Two strains carried subtypes Stx2h and Stx2o, which have only one previous report. Two strains carried a previously undescribed subtype, Stx2j.

The global population structure and evolutionary history of the acquisition of major virulence factor-encoding genetic elements in Shiga toxin-producing Escherichia coli O121:H19

Shiga toxin (Stx)-producing Escherichia coli (STEC) are foodborne pathogens causing serious diseases, such as haemorrhagic colitis and haemolytic uraemic syndrome. Although O157:H7 STEC strains have been the most prevalent, incidences of STEC infections by several other serotypes have recently increased. O121:H19 STEC is one of these major non-O157 STECs, but systematic whole genome sequence (WGS) analyses have not yet been conducted on this STEC. Here, we performed a global WGS analysis of 638 O121:H19 strains, including 143 sequenced in this study, and a detailed comparison of 11 complete genomes, including four obtained in this study. By serotype-wide WGS analysis, we found that O121:H19 strains were divided into four lineages, including major and second major lineages (named L1 and L3, respectively), and that the locus of enterocyte effacement (LEE) encoding a type III secretion system (T3SS) was acquired by the common ancestor of O121:H19. Analyses of 11 complete genomes belonging to L1 or L3 revealed remarkable interlineage differences in the prophage pool and prophage-encoded T3SS effector repertoire, independent acquisition of virulence plasmids by the two lineages, and high conservation in the prophage repertoire, including that for Stx2a phages in lineage L1. Further sequence determination of complete Stx2a phage genomes of 49 strains confirmed that Stx2a phages in lineage L1 are highly conserved short-tailed phages, while those in lineage L3 are long-tailed lambda-like phages with notable genomic diversity, suggesting that an Stx2a phage was acquired by the common ancestor of L1 and has been stably maintained. Consistent with these genomic features of Stx2a phages, most lineage L1 strains produced much higher levels of Stx2a than lineage L3 strains. Altogether, this study provides a global phylogenetic overview of O121:H19 STEC and shows the interlineage genomic differences and the highly conserved genomic features of the major lineage within this serotype of STEC.

Distribution of virulence factors, antimicrobial resistance genes and phylogenetic relatedness among Shiga toxin-producing Escherichia coli serogroup O91 from human infections

Shiga toxin-producing Escherichia coli (STEC) belonging to the serogroup O91 are among the most common non-O157 STEC serogroups associated with human illness in Europe. This study aimed to analyse the virulence factors, antimicrobial resistance genes and phylogenetic relatedness among 48 clinical STEC O91 isolates collected during 2003-2019 in Switzerland. The isolates were subjected to whole genome sequencing using short-read sequencing technologies and a subset of isolates additionally to long-read sequencing. They belonged to O91:H10 (n=6), O91:H14 (n=40), and O91:H21 (n=2). Multilocus sequence typing showed that the O91:H10 isolates all belonged to sequence type (ST)641, while the O91:H14 isolates were assigned to ST33, ST9700, or were non-typeable. Both O91:H21 isolates belonged to ST442. Shiga toxin gene stx1a was the most common Shiga toxin gene subtype among the isolates, followed by stx2b, stx2d and stx2a. All isolates were LEE-negative and carried one or two copies of the IrgA adhesin gene iha. In a subset of long-read sequenced isolates, modules of the Locus of Adhesion and Autoaggregation pathogenicity island (LAA-PAI) carrying iha and other genes such as hes, lesP or agn43 were identified. A large proportion of STEC O91:H14 carried the subtilase cytotoxin gene subA, colicin genes (cba, cea, cib and cma) or microcin genes (mcmA, mchB, mchC and mchF). STEC O91:H14 were further distinguished from STEC O91:H10/H21 by one or more virulence factors found in extraintestinal pathogenic E. coli (ExPEC), including hlyF, iucC/iutA, kpsE and traT. The hlyF gene was identified on a novel mosaic plasmid that was unrelated to hlyF+ plasmids described previously in STEC. Core genome phylogenetic analysis revealed that STEC O91:H10 and STEC O91:H21 were clonally conserved, whereas STEC O91:H14 were clonally diverse. Among three STEC O91:H14 isolates, a number of resistance genes were identified, including genes that mediate resistance to aminoglycosides (aadA, aadA2, aadA9, aadA23, aph(3'')-Ib and aph(6)-Id), chloramphenicol (cmlA), sulphonamides (sul2 and sul3), and trimethoprim (drfA12). Our data contribute to understanding the genetic diversity and differing levels of virulence potential within the STEC O91 serogroup.

Emergence of New ST301 Shiga Toxin-Producing Escherichia coli Clones Harboring Extra-Intestinal Virulence Traits in Europe

O80:H2 enterohemorrhagic Escherichia coli (EHEC) of sequence type ST301 is one of the main serotypes causing European hemolytic and uremic syndrome, but also invasive infections, due to extra-intestinal virulence factors (VFs). Here, we determined whether other such heteropathotypes exist among ST301. EnteroBase was screened for ST301 strains that were included in a general SNP-phylogeny. French strains belonging to a new heteropathotype clone were sequenced. ST, hierarchical clusters (HC), serotype, resistome, and virulome were determined using EnteroBase, the CGE website, and local BLAST. The ST301 general phylogeny shows two groups. Group A (n = 25) is mainly composed of enteropathogenic E. coli, whereas group B (n = 55) includes mostly EHEC. Three serotypes, O186:H2, O45:H2 and O55:H9, share the same virulome as one of the O80:H2 sub-clones from which they derive subsequent O-antigen switches. The O55:H9 clone, mainly present in France (n = 29), as well as in the UK (n = 5) and Germany (n = 1), has a low background of genetic diversity (four HC20), although it has three Stx subtypes, an H-antigen switch, and genes encoding the major extra-intestinal VF yersiniabactin, and extended-spectrum beta-lactamases. Diverse heteropathotype clones genetically close to the O80:H2 clone are present among the ST301, requiring close European monitoring, especially the virulent O55:H9 clone.

Insights into the genome architecture and evolution of Shiga toxin encoding bacteriophages of Escherichia coli

Background

A total of 179 Shiga toxin-producing Escherichia coli (STEC) complete genomes were analyzed in terms of serotypes, prophage coding regions, and stx gene variants and their distribution. We further examined the genetic diversity of Stx-converting phage genomes (Stx phages), focusing on the lysis-lysogeny decision and lytic cassettes.

Results

We show that most STEC isolates belong to non-O157 serotypes (73 %), regardless the sources and geographical regions. While the majority of STEC genomes contain a single stx gene (61 %), strains containing two (35 %), three (3 %) and four (1 %) stx genes were also found, being stx2 the most prevalent gene variant. Their location is exclusively found in intact prophage regions, indicating that they are phage-borne. We further demonstrate that Stx phages can be grouped into four clusters (A, B, C and D), three subclusters (A1, A2 and A3) and one singleton, based on their shared gene content. This cluster distribution is in good agreement with their predicted virion morphologies. Stx phage genomes are highly diverse with a vast number of 1,838 gene phamilies (phams) of related sequences (of which 677 are orphams i.e. unique genes) and, although having high mosaicism, they are generally organized into three major transcripts. While the mechanisms that guide lysis–lysogeny decision are complex, there is a strong selective pressure to maintain the stx genes location close to the lytic cassette composed of predicted SAR-endolysin and pin-holin lytic proteins. The evolution of STEC Stx phages seems to be strongly related to acquiring genetic material, probably from horizontal gene transfer events.

Conclusions

This work provides novel insights on the genetic structure of Stx phages, showing a high genetic diversity throughout the genomes, where the various lysis-lysogeny regulatory systems are in contrast with an uncommon, but conserved, lytic system always adjacent to stx genes.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12864-021-07685-0.

Genetic characterization of Shigatoxigenic and enteropathogenic Escherichia coli O80:H2 from diarrhoeic and septicaemic calves and relatedness to human Shigatoxigenic E. coli O80:H2

AIM

The purpose of this work was to identify and genetically characterize enterohemorrhagic Escherichia coli (EHEC) and enteropathogenic E. coli (EPEC) O80:H2 from diarrhoeic and septicaemic calves in Belgium and to comparing them with human EHEC after whole genome sequencing.

METHODS AND RESULTS

Ten EHEC and 21 EPEC O80 identified by PCR between 2009 and 2018 from faeces, intestinal content and a kidney of diarrhoeic or septicaemic calves were genome sequenced and compared to 19 human EHEC identified between 2008 and 2019. They all belonged to the O80:H2 serotype and ST301, harboured the eaeξ gene, and 23 of the 29 EHEC contained the stx2d gene. Phylogenetically, they were distributed in two major sub-lineages: one comprised a majority of bovine EPEC whereas the second one comprised a majority of stx2d bovine and human EHEC.

CONCLUSIONS

Not only EPEC but also EHEC O80:H2 are present in diarrhoeic and septicaemic calves in Belgium and are genetically related to human EHEC.

SIGNIFICANCE AND IMPACT OF THE STUDY

These findings support the need to assess cattle as potential source of contamination of humans by EHEC O80:H2 and to understand the evolution of bovine and human EHEC and EPEC O80:H2.

Multiple-Locus Variable-Number Tandem Repeat Analysis Scheme for Non-O157 Shiga Toxin-Producing Escherichia coli: Focus on Serogroups O103, O121, O145, O165, and O91

Non-O157 Shiga toxin-producing Escherichia coli (STEC) infections are a growing concern for public health. The number of sporadic cases and outbreaks of non-O157 STEC infections have increased in recent years. Molecular subtyping is an essential tool that allows high-resolution and rapid differentiation of isolates, identification of case clusters, and detection of outbreak clusters. Multiple-locus variable-number tandem repeat analysis (MLVA) is one of the most useful typing methods for differentiating isolates that cause foodborne diseases. In Japan, serogroups O26, O111, O103, O121, O145, O165, and O91 have been frequently isolated or associated with severe cases of non-O157 STEC infections. In this study, we designed an MLVA scheme (MLVA43) for serogroups O103, O121, O145, O165, and O91 by adding 26 new loci to an MLVA scheme (MLVA17) previously developed by our group for serogroups O157, O26, and O111 using 17 loci. We found that the discriminatory power of MLVA43 was comparable to that of pulsed-field gel electrophoresis (PFGE) for serogroups O103, O145, O165, and O91, and superior to that of PFGE for O121. MLVA43 identified more profiles than did MLVA17, except for serogroup O111 with 707 isolates. The MLVA43 scheme will enable rapid detection of outbreak clusters, which will aid in implementing rapid control measures against non-O157 STEC infections.

Is Shiga Toxin-Producing Escherichia coli O45 No Longer a Food Safety Threat? The Danger is Still Out There

Many Shiga toxin-producing Escherichia coli (STEC) strains, including the serogroups of O157 and most of the top six non-O157 serotypes, are frequently associated with foodborne outbreaks. Therefore, they have been extensively studied using next-generation sequencing technology. However, related information regarding STEC O45 strains is scarce. In this study, three environmental E. coli O45:H16 strains (RM11911, RM13745, and RM13752) and one clinical E. coli O45:H2 strain (SJ7) were sequenced and used to characterize virulence factors using two reference E. coli O45:H2 strains of clinical origin. Subsequently, whole-genome-based phylogenetic analysis was conducted for the six STEC O45 strains and nine other reference STEC genomes, in order to evaluate their evolutionary relationship. The results show that one locus of enterocyte effacement pathogenicity island was found in all three STEC O45:H2 strains, but not in the STEC O45:H16 strains. Additionally, E. coli O45:H2 strains were evolutionarily close to E. coli O103:H2 strains, sharing high homology in terms of virulence factors, such as Stx prophages, but were distinct from E. coli O45:H16 strains. The findings show that E. coli O45:H2 may be as virulent as E. coli O103:H2, which is frequently associated with severe illness and can provide genomic evidence to facilitate STEC surveillance.

Differential dynamics and impacts of prophages and plasmids on the pangenome and virulence factor repertoires of Shiga toxin-producing Escherichia coli O145:H28

Phages and plasmids play important roles in bacterial evolution and diversification. Although many draft genomes have been generated, phage and plasmid genomes are usually fragmented, limiting our understanding of their dynamics. Here, we performed a systematic analysis of 239 draft genomes and 7 complete genomes of Shiga toxin (Stx)-producing Escherichia coli O145:H28, the major virulence factors of which are encoded by prophages (PPs) or plasmids. The results indicated that PPs are more stably maintained than plasmids. A set of ancestrally acquired PPs was well conserved, while various PPs, including Stx phages, were acquired by multiple sublineages. In contrast, gains and losses of a wide range of plasmids have frequently occurred across the O145:H28 lineage, and only the virulence plasmid was well conserved. The different dynamics of PPs and plasmids have differentially impacted the pangenome of O145:H28, with high proportions of PP- and plasmid-associated genes in the variably present and rare gene fractions, respectively. The dynamics of PPs and plasmids have also strongly impacted virulence gene repertoires, such as the highly variable distribution of stx genes and the high conservation of a set of type III secretion effectors, which probably represents the core effectors of O145:H28 and the genes on the virulence plasmid in the entire O145:H28 population. These results provide detailed insights into the dynamics of PPs and plasmids, and show the application of genomic analyses using a large set of draft genomes and appropriately selected complete genomes.