- Research article
- Open Access
Genome-wide mosaicism within Mycobacterium abscessus: evolutionary and epidemiological implications
- Guillaume Sapriel1, 2, 3,
- Julie Konjek†1, 4,
- Mickael Orgeur†5,
- Laurent Bouri1,
- Lise Frézal6,
- Anne-Laure Roux7,
- Emilie Dumas1,
- Roland Brosch5,
- Christiane Bouchier7,
- Sylvain Brisse8,
- Mathias Vandenbogaert8,
- Jean-Michel Thiberge8,
- Valérie Caro8,
- Yun Fong Ngeow9,
- Joon Liang Tan9,
- Jean-Louis Herrmann1, 10,
- Jean-Louis Gaillard1, 4,
- Beate Heym1, 4 and
- Thierry Wirth11, 12Email author
© Sapriel et al. 2016
- Received: 3 February 2015
- Accepted: 8 February 2016
- Published: 17 February 2016
In mycobacteria, conjugation differs from the canonical Hfr model, but is still poorly understood. Here, we quantified this evolutionary processe in a natural mycobacterial population, taking advantage of a large clinical strain collection of the emerging pathogen Mycobacterium abscessus (MAB).
Multilocus sequence typing confirmed the existence of three M. abscessus subspecies, and unravelled extensive allelic exchange between them. Furthermore, an asymmetrical gene flow occurring between these main lineages was detected, resulting in highly admixed strains. Intriguingly, these mosaic strains were significantly associated with cystic fibrosis patients with lung infections or chronic colonization. Genome sequencing of those hybrid strains confirmed that half of their genomic content was remodelled in large genomic blocks, leading to original tri-modal ‘patchwork’ architecture. One of these hybrid strains acquired a locus conferring inducible macrolide resistance, and a large genomic insertion from a slowly growing pathogenic mycobacteria, suggesting an adaptive gene transfer. This atypical genomic architecture of the highly recombinogenic strains is consistent with the distributive conjugal transfer (DCT) observed in M. smegmatis. Intriguingly, no known DCT function was found in M. abscessus chromosome, however, a p-RAW-like genetic element was detected in one of the highly admixed strains.
Taken together, our results strongly suggest that MAB evolution is sporadically punctuated by dramatic genome wide remodelling events. These findings might have far reaching epidemiological consequences for emerging mycobacterial pathogens survey in the context of increasing numbers of rapidly growing mycobacteria and M. tuberculosis co-infections.
- Horizontal Gene Transfer
- Cystic Fibrosis Patient
- Genomic Island
- Multi Locus Sequence Typing
- Mycobacterium Abscessus
Clonal evolution was a long lasting paradigm in mycobacterial research with the highly clonal flagship of the genus, Mycobacterium tuberculosis. Ultimately the rule turned-out to be the exception and true clonal species are rather rare; the most representative members except M. tuberculosis are Yersinia pestis, Salmonella typhi and Burkholderia mallei. However, there is increasing evidence that horizontal gene transfer (HGT) and homologous DNA recombination play an important role in the evolution of smooth tubercle bacilli [1, 2] and M. tuberculosis strains [3–5]. Gene transfer networks mostly involving genes related to niche change and antibiotic resistance are significantly shaping the adaptive landscape of pathogenic mycobacteria, therefore deciphering the mechanisms behind these empirical observations becomes mandatory. Moreover, knowledge of such mechanisms can shed new light on mycobacterial evolution from saprophytic/commensal organisms to opportunistic or specialized, highly persisting pathogens [3, 6]. However, knowledge about HGT and homologous recombination mechanisms in mycobacteria are scarce. Studies on the saprophytic laboratory model Mycobacterium smegmatis showed that chromosomal DNA transfer is mechanistically different from classical Hfr chromosomal DNA transfer, with multiple and wide-spread transfer initiations events from a donor chromosome [7, 8]. This process, called distributive conjugal transfer, creates extensive genome-wide mosaicism within individual transconjugants that generates large-scale sibling diversity conferring the evolutionary benefits of sexual reproduction in an asexual organism [9, 10]. The chromosomal region involved in this unique conjugation mechanism is the ESX-1 secretion system , which is also involved in virulence in M. tuberculosis .
In the present study, our interest focused on M. abscessus (MAB). This mycobacteria is an excellent model to study HGT, homologous recombination and their contribution to pathogenicity in mycobacteria. MAB is an emerging opportunistic pathogen, able to cause lung diseases to immunocompetent individuals and that shares a number of characteristics with M. tuberculosis, such as the ability to induce granulomatomatous lesions with epithelioid giant cells, caseous necrosis, and silent persistence for decades within host . Since the late 1990’s, MAB has been increasingly recovered from patients with cystic fibrosis (CF) in Europe, Asia and North America [14–16]. Together with M. avium, MAB represents the most commonly isolated non-tuberculous mycobacteria (NTM) from CF lung patients. Reports show that MAB isolated from CF patients can account for up to 56 % of all isolated NTM . Compared with other NTM causing pulmonary diseases, MAB can be considered as the most pathogenic, since all reports show that this NTM has the highest rate of genuine ‘clinically relevant’ infections in CF patients according to criteria established by the American Thoracic Society (ATS) [18–22]. Antibiotic resistance is also a major factor in the high rate of treatment failure for MAB pulmonary diseases (20 to 52 %) [23–25]. MAB lung infections cause decline of lung function , and dissemination of the infectious agent, eventually leading to death [26, 27]. Moreover, MAB is resistant to nearly all antibiotics, including first-line antitubercular drugs , and the few active antibiotics only seem to have a bacteriostatic effect . Taken together, these features make MAB an emerging pathogen under close surveillance. Furthermore, MAB is also satisfying population studies criteria, since large isolate collections are available, with systematically documented clinical profiles and cohort studies, especially in the case of cystic fibrosis (CF) patients .
Another characteristic making MAB a very relevant model for genetic exchange study in mycobacteria is the fact that it harbors various phenotypes such as rough and smooth morphotypes , macrolid resistance , and anaerobic growth . MAB encompasses a large genetic diversity  that is markedly associated with different prevalence [34–36], specific involvement in outbreaks , distinct virulence  and contrasting clinical outcomes .
Recent genomic studies comparing the genomes of 40 strains from the MAB unravelled a large repertoire of accessory genes, suggesting extensive genetic acquisition capacities and high evolutionary potential for this species by HGT . This trend was confirmed with the publication of a reference genome sequence obtained by Sanger method, showing that some virulence genes might have been acquired by HGT from non-mycobacterial species sharing a similar ecological niche . Moreover, MAB is subdivided into at least two recognized subspecies: Mycobacterium abscessus subsp. abscessus and M. abscessus subsp. bolletii [42, 43], making this species an excellent model for studying intra-specific homologous recombination. Multi Locus Sequence typing (MLST) studies showed that some clinical isolates have a composite genetic pattern with housekeeping genes corresponding to different subspecies, suggesting that homologous recombination occurs readily within the MAB [34, 35, 44, 45]. However, unlike M. smegmatis, sequence analysis of 14 MAB genomes showed that no ESX-1 orthologous system is encoded within this species, whereas ESX-3 and ESX-4 secretion systems are present , raising the question whether alternative DCT systems might exist. Studying mycobacteria-specific HGT dynamics and the involved cellular machinery might definitively improve our understanding of the subspecies border delimitations and the amount of gene flow occurring within the MAB.
Although HGT (exogenous insertions as well as inter-strain homologous recombination) seem to be extensive in MAB strains, no quantitative data are available regarding the genetic flux between specific MAB subspecies, and the genetic architecture (i.e., location, distribution, and extent) of these genetic exchanges on the genome are unknown. Thus, using an MLST approach, and taking advantage of a set of 280 clinical strains, our goal was to extensively characterize the genetic exchanges occurring between MAB subspecies. This approach led us to identify a sub-population of highly admixed strains. Furthermore, using comparative genomics, we succeeded in generating comprehensive recombination cartography for some of these admixed strains. This highly admixed sub-population was then further investigated for virulence by using clinical records. Finally, in-depth genomic analysis was performed in order to identify putative specific DCT functions.
Phylogenetic signal and analyses
Prior conducting any phylogenetic inferences based on the 7 MLST gene fragments (argH, Cya, glpK, gnd, murC, pta, and purH), we inferred the quality of the phylogenetic information contained in these sequences by plotting transition and transversion rates as a function of genetic distances (Additional file 1: Figure S1). This graph shows that neither transitions nor transversions are saturated but, rather, both rise linearly with increasing genetic distance. We then measured substitution saturation using the Xia index  for all three codon positions. The observed I ss.c value of 0.808 was significantly higher than the I ss value of 0.019, thus confirming that little saturation occurs at these sites.
Population genetics and recombination
After having attributed each sequence type (ST) to a given subspecies, we determined the percentage of exogenous SNPs for each ST (Fig. 2d). Subspecies belonging to M. massiliense and M. abscessus strains display the highest proportion of foreign SNPs, whereas M. bolletii strains are far more homogenous. Moreover, allelic flux going from M. massiliense to M. bolletii was significantly lower than other allelic exchanges (Fisher test P < 0.01. Fig. 2d). Consistently, mosaic strains proportion is significantly higher for M. massiliense STs than for M. bolletii (Fisher test P < 0.05. Fig. 2e). Although they should be confirmed by whole genomic survey, these results suggest that allelic fluxes between the three subspecies are not homogenous. They indicate asymmetrical gene flow between subspecies, especially between M. massiliense and M. bollettii.
Population estimates of mutation rates (θ) and recombination rates (ρ) per base
r / μ
CI 95 %
A side effect of the allelic flux within MAB concerns the suitability of rpoB for subspecies identification. Indeed, there is no reason that this gene will escape interlineage homologous recombination and consequently using it as a diagnostic marker for species determination within the MAB might be misleading (Additional file 4: Figure S4A). According to STRUCTURE assignments, and consistent with our previous result emphasizing M. massiliense high recombination rates, the false group identification rate is about 10 % within MAB and climbs up to 20 % for strains belonging to M. massiliense (Additional file 4: Figure S4B).
Clinical symptoms of infection and mosaicism
Clinical information was collected from 102 different MAB isolates from infected patients with well-documented geographic origins and clinical background (See Additional file 5: Table S1). The patients were divided into sub-groups according to the type of infection (non-pulmonary vs pulmonary infections), and within pulmonary infections patients were divided according to the underlying clinical context (immunocompromised patient, respiratory underlying disease, non-respiratory chronic disease, and cystic fibrosis). Within this sample, CF patients were gathered from a former multi-center study, and were monitored on several years (1996–2009), allowing the identification of chronic MAB colonisation. These patients were also precisely identified according to the clinical manifestation of non-tubercular lung disease (NTMLD) showing clinical symptoms suggestive of lung infection. An exogenous allele acquisition index (inter subspecies) was computed for each isolate, using STRUCTURE and rpoB results to determine if at least one of the eight sequenced genes came from exogenous subspecies. The results show that strains belonging to mosaic STs are over-represented in CF patients with MAB infection or chronic lung colonisation (Fig. 3c). Whereas CF patients with MAB infection or chronic colonisation had a majority of admixed MAB isolates, most non-CF patients were infected with low proportion of admixed MAB strains (except in the case of patients with non-pulmonary chronic disease). These proportions differ significantly (Fisher exact test P = 0.038 for CF patients with MAB chronic colonisation, and P = 0.034 for CF patients with MAB lung infection). More specifically, the proportion of admixed MAB significantly differed between patients with non-respiratory infections, as compared with CF patients with MAB infection or chronic colonisation (Fisher exact test P = 0.03 for both patient category). The most marked difference was observed for patients having MAB pulmonary infection with non-pulmonary underlying chronic disease. MAB associated with those patients very significantly differed from other stains (Fisher exact test P = 0.008). In order to exclude potential cross-contamination effect, the same calculations were performed with only single STs, and the differences observed between the most extreme distribution was still significant (for patients having pulmonary infection with non-pulmonary underlying chronic disease : Fisher exact test P = 0.029 as compared to patients with Non-pulmonary infections, and P = 0.039 as compared to all other patients). Those results suggest that different MAB populations, showing contrasted patterns of genetic admixture, are preferentially associated with some specific clinical profiles (depending on infection types and patient clinical background).
HGT mapping and comparative genomics
Core, variable, and strain specific open reading frames (ORFs)
Strain spe. ORFs
Core ORFs (%)
Var ORFs (%)
Strain spe. ORFs (%)
Mycobacterium abscessus ATCC 19977
Mycobacterium massiliense CIP 108297 T
Mycobacterium bolletii CIP 108541 T
Mycobacterium massiliense strain M139
Mycobacterium abscessus strain 137
Mycobacterium massiliense strain 23
Large recombination domains (up to 93 kb) are also found all along the genomes, and may account for up to a quarter of the genomic content. Interestingly, chromosomal mapping domains that are identical to each of the reference strains show a non-random distribution. These domains are clustered together into large continuous regions that are associated with one of the three reference strains (Fig. 4 and Additional file 7: Figure S6). For example, in strain M139, the length of the largest domain similar to M. abscessus was 53 kb (median length 5.4 kb); the length of the largest domain similar to M. bolletii was 48 kb (median length 4 kb) and the length of the longest domain similar to none of the reference strains was 61.5 kb (median length 4 kb). Worth mentioning, our results confirm that strain M139 belongs to the M. massiliense subspecies and harbours an M. abscessus subspecies erm (41) allele (usually conferring inducible macrolide resistance) [50, 51]. This erm(41) element is embedded in a large continuous 60 kb genomic region also clearly assigned to an M. abscessus subspecies genetic background (Additional file 8: Figure S7A). Interestingly, the same M139 highly mosaic strain also contained a large non-aligned contig containing a locus highly similar to the M. marinum p-RAW conjugative mega-plasmid (Additional file 8: Figure S7B) [39, 52]. Thus, our results show that MAB population is structured into three sub-species, and also contains some unclassifiable mosaic strains. These unclassifiable individuals are highly plastic (that have almost half of their genomic content remodelled by HGT) with a pattern that strikingly resembles DCT genome mosaicism reported in M. smegmatis and M. canettii .
Here, we sampled and analysed a representative and balanced collection of strains that represents the MAB diversity, both at phylogenetic and epidemiological levels, and from different geographical sources. Our results obtained from MLST analyses combined with whole genome sequence analysis of some representative strains clearly confirm the existence of 3 subspecies and therefore cast some doubts on the classification proposed by Leao and colleagues . These authors suggested to group M. bollettii and M. massiliense together into a subspecies named M. abscessus subspecies bolletii comb. nov.
On the other hand, according to the large amounts of putative gene flow detected by STRUCTURE and LDHAT (Table 1), the three entities described here do not fully fit the genospecies described by Drancourt and colleagues . Indeed, our MLST data suggest that homologous recombination in the MAB has been frequent enough to generate about one third of strains within the species with traces of mixed ancestries (Fig. 2b), from which another third (so-called admixed strains) have at least 20 % of their gene content from a foreign source. Even more striking is the fact that significant homologous recombination was detected both within and between subspecies (Table 1). This means that the MAB does not fit into a clonal framework and that the amount of genetic exchange detected here resembles the one reported in a previous study on Escherichia coli . However, the data show that the three subspecies do not behave the same way. Indeed, M. bolletii is clearly less introgressed than the two other subspecies, resulting in a relatively homogenous gene pool that might result from a distinct or isolated ecological niche. It is tempting to link this observation with the rather low prevalence of M. bolletii in cystic fibrosis  and other chronic pulmonary infections . Moreover, the limited genic repertoire of this subspecies combined with its mostly clonal propagation will limit its adaptive landscape in a clinical setting.
The situation is dramatically different for M. abscessus and M. massiliense, where homologous recombination is prevalent. For example, in M. abscessus, each nucleotidic change has nearly seven more chances to be generated by recombination than by mutation. There is accumulating evidence that recombinogenic species display higher virulence profiles and plasticity than clonal ones [54, 55]. This situation is encountered in the MAB, where isolates gathered from CF patient with record of clinical symptoms have exchanged significantly more alleles between subspecies than isolates from patients with other clinical profiles (Fig. 3c). This observation might also be linked to the fact that the majority of MAB infections in CF patients are silent, whereas only 10 % of the isolates are associated with clinical symptoms of pulmonary disease (Roux et al. in press), which is reminiscent of the 10 % of hyper-mosaic strain observed. Taken together, these observations show a link between chronic lung colonization and mosaicism in MAB. This could be driven by different scenarios: MAB mosaic strains might have acquired genes or alleles conferring greater virulence profile and/or lung colonization ability, alternatively the specific conditions associated with pulmonary tissue colonisation (such as host defence, or drug treatments) may submit colonizing strain to environmental stresses that favour genetic exchanges.
From a clinical point of view, our study does not provide any trend or information that favours one type of subspecies in a given clinical situation. Furthermore, soft tissue infections, lung diseases and systemic diseases were not preferentially associated with any subspecies. This absence of clinical and geographical correlations confirms the opportunistic, accidental nature of the infections, most likely from an environmental source. In terms of clinical diagnostics, we clearly illustrate the lack of power of rpoB typing that culminates at 20 % failure rates in the M. massiliense subspecies. Accordingly, molecular diagnostics might definitely profit from a multilocus typing scheme.
Nucleotide BLAST results of admixed genomes genomic islands
Genomic island Start position
Genomic island End position
First nucleotide blast resulta
Mycobacterium massiliense strain M139
Mycobacterium intracellulare MOTT-02, complete genomeb
1.12E + 005
Mycobacterium massiliense strain M139
Mycobacterium sp. JDM601, complete genome b
Mycobacterium massiliense strain M139
Mycobacterium smegmatis JS623, complete genome
Mycobacterium massiliense strain M139
Mycobacterium chubuense NBB4, complete genome
Mycobacterium massiliense strain 23
Mycobacterium avium 104, complete genome
Mycobacterium massiliense strain 23
Mycobacterium gilvum Spyr1, complete genome
Mycobacterium abscessus strain 137
Mycobacterium intracellulare MOTT-02, complete genomeb
1.07E + 005
Mycobacterium abscessus strain 137
Mycobacterium sp. JLS, complete genome
Mycobacterium abscessus strain 137
Mycobacterium ulcerans Agy99, complete genomeb
The genetic architecture of the admixed strains advocates for rather rare but massive genetic exchanges between MAB subspecies and only hardly fits with long-lasting and regular gene flow that would lead to highly scattered patterns disrupting the genome-wide mosaicism into a gene-wide mosaicism. In the context of an absence of ESX1 loci in our MAB collection and with no evidence that ESX3 or ESX4 play a role in M. smegmatis DCT (which also contains those two types of ESX), the genetic elements involved in MAB mosaicism remain unresolved. However, the two extrachromosomal elements carrying ESX/typeIV systems detected in some MAB strains share many traits with the novel p-RAW plasmids discovered in M. marinum and other SGM [52, 57] (Additional file 8: Figure S7B). Recent investigations also showed that other similar ESX/typeIV systems exist in mycobacteria and belong to a quite large and diversified family. Taken together, these results raise the hypothesis that, besides the genomic ESX1-driven DCT, other extrachromosomal p-RAW-like ESX elements might be involved in mycobacterial conjugation, and allow subsequent DCT. We are aware that this novel hypothetical form of plasmid-driven conjugation must be tested in experimental F1-generation transconjugant experiments to evaluate the evolutionary and pathogenic potential behind this system in MAB.
Admixed populations of MAB seem to display higher abilities for colonizing lungs of CF patients. On the other hand, we cannot exclude that long-term lung colonization might also favour MAB genetic admixture and HGT in this very specific ecological niche. Our study also strongly suggests that pRAW-like extra-chromosomal genetic elements might be responsible for the massive genomic exchanges observed in MAB, and are reminiscent of those observed in distributive conjugal transfer coded on chromosomal ESX1 system in M. smegmatis. Therefore, their contribution to HGT in mycobacterial evolution and pathogenicity should be assessed in a general context of increasing MAB lung infections and MAB/TB co-infections.
Bacterial strains collection and sequence dataset
A total of 280 strains belonging to the MAB were chosen from multiple clinical sources in diverse geographical areas in an attempt to study the genetic diversity of this bacterial species from an evolutionary perspective (Additional file 5: Table S1). In order to verify that all strains belonged to MAB, RpoB typing was performed, and phylogenetic trees of each of the seven housekeeping genes used for MLST analysis (argH, Cya, glpK, gnd, murC, pta, and purH), were carefully checked (see Additional file 9: Figure S8). We also assessed that the associated housekeeping genes were under strong purifying selection (Additional file 10: Figure S9). To estimate the level of gene conservation, pairwise dN/dS ratio ω (dN: non-synonymous mutation substitution rate, dS: synonymous mutation substitution rate) were calculated using the program CODEML provided by the PAML (Phylogenetic Analyses by Maximum Likelihood) package version 4 . Nucleotidic sequences have been aligned using TRANSLATORX  guided by protein sequence alignments obtained using M-COFFEE .
Clinical information was collected from 102 different MAB isolates from different infected patients with precise documented geographic origins and clinical background (See Additional file 5: Table S1). In order to avoid samples biases, we only used one isolate per patient and per reported outbreak. These patients were divided into sub-groups according to the type of infected tissue (non-pulmonary vs pulmonary infections). Within pulmonary infections patients were divided according to the underlying clinical context (immunocompromised patient, respiratory underlying disease, non-respiratory chronic disease, and cystic fibrosis). CF patients clinical profile were precisely documented from a French multicenter cohort study . All the patients, or their parents if they were children, gave their informed consent. Data were retrieved from the French CF registry (CNIL authorisation No 1,202,233), and an internal review board approved the study. Samples were analyzed for NTM identification at each center, using approved techniques and all data analysed were anonymized. CF patients were considered as infected if MAB was associated with “nontuberculous mycobacterial lung disease” (NTMLD), if the cases: i) fulfilled the bacteriological American Thoracic Society criteria for mycobacterial lung infections , and ii) presented clinical (e.g., functional deterioration such as fever, asthenia and emaciation) and/or radiographic signs of mycobacterial disease. Each CF patient was periodically investigated for pulmonary MAB isolation between 1996 and 2009. If MAB was no more identified after previous positive tests, and without any anti-MAB treatment, this was considered as a ‘transient colonisation’. If MAB was isolated at each investigation after first isolation on a minimal period of 5 years, this was considered as a ‘chronic colonisation’.
Gene fragments sequencing
Eight gene fragments were amplified and sequenced from all isolates using the primers and PCR protocols presented in Macheras et al. 2011. Both strands were sequenced using an Applied Biosystems Prism 3700 automated sequencer with dRhodamine-labeled terminators (PE Applied Biosystems). Sequences were aligned and trimmed using SEQLAB and PILEUP (Wisconsin Package 9.1, GCG, Madison, WI) and then concatenated. All sequence type profiles and nucleotide sequences are publicly available at (http://bigsdb.web.pasteur.fr/mycoabscessus/mycoabscessus.html).
MLST and minimum spanning-trees
In order to define the relationships between strains at the microevolution level, we performed allelic profile-based comparisons using a minimal spanning tree (MST) analysis with the BIONUMERICS v5.10 software (Applied-Maths, Sint Maartens-Latem, Belgium). The minimal spanning tree is calculated by Prim’s algorithm, modified to choose between otherwise equivalent, alternative subtrees at each step by implementing priority rules that incorporate aspects of the EBURST algorithm . The highest priority is given to STs with the largest numbers of single locus variants. Any ties were resolved by choosing the ST (or a random ST) with the largest number of isolates. The first node in the network is the ST with the highest priority according to these rules and subsequent links are chosen by a recursive strategy. ST complexes were defined as containing at least three STs, with links of one or two shared alleles. The graphical representation displays the quantitative relationships between STs and ST complexes, measured as the number of shared alleles, by lines of different thickness and type.
In order to gain an overview of the phylogenetic signal, we plotted pairwise transition and transversion distances against the total genetic distances using the DAMBE software package , and we also tested our set of molecular sequences for substitution saturation. The phylogenetic signal of the dataset was also investigated with the likelihood mapping method implemented in TREE-PUZZLE  by analysing 10,000 random quartets. This method proceeds by evaluating, using maximum likelihood, groups of four randomly chosen sequences (quartets). The three possible unrooted tree topologies, for each quartet, are weighted and the posterior weights are then plotted using triangular coordinates, such that each corner represents a fully resolved tree topology. Therefore the resulting distribution of the points shows whether the data are suitable for a phylogenetic reconstruction, or not. The best-fit model of DNA substitution and the parameter estimates used for tree reconstruction were chosen by performing hierarchical likelihood ratio tests implemented in JMODELTEST 2.1.3 . Phylogenetic trees were estimated for each data set with PHYML incorporating the best-fit model of evolution. Alternatively, we also implement the Bayesian Markov chain Monte Carlo (MCMC) method available in the BEAST 1.8.1 package  to generate a phylogeny of MAB. The general time reversible (GTR) substitution model was implemented under a constant population size scenario and five independent runs were generated. Convergence was then evaluated with ESS values and trace plots were explored with the software TRACER 1.5. Runs were combined using LOGCOMBINER and trees were plotted using FIGTREE v1.3.2 and DENSITREE 2.0.1. Split decomposition analyses were performed with SPLITSTREE, version 4 , by using LogDet distances, equal edge lengths, and 1000 bootstrap replicates.
Population genetic analyses
We used the linkage model in STRUCTURE  to identify groups with distinct allele frequencies [54, 68, 69]. This procedure assigns a probability of ancestry for each polymorphic nucleotide for a given number of groups, K, and also estimates q, the combined probability of ancestry from each of the K groups for each individual isolate. As given by the Evanno’s test , we chose three groups for this report because repeated analyses (200,000 iterations following a burn-in period of 80,000 iterations) with K between 1 and 10 showed that the model probability increased dramatically between K = 2 and K = 3 and only slowly thereafter. A cut-off value of q ≥ 0.80 was then used to assign individual isolates to one of the three groups that largely matched the classical nomenclature. Unassigned isolates were designated as “hybrid” strains.
Recombination and mutation
We tested for recombination within the 3 subspecies for all loci independently, as well as for the concatenated MLST loci using the software LDHAT v2.2 . LDHAT employs a coalescent-based method to estimate the population-scaled mutation (θ = 2N e μ) and recombination (ρ = 2N e r) rates, where N e is the effective population size, r the rate at which recombination events separate adjacent nucleotides and μ is the mutation rate per nucleotide. The ratio r/μ were calculated as (ρ/L)/θ, where L is the gene length (sequence length). This r/μ ratio ranges from 0, which indicates full clonal reproduction, to > > 1, which is expected under free recombination. Significance of the evidence for recombination was tested using non-parametric, permutation-based tests implemented in LDHAT (Lkmax and G4 tests). To avoid strains overrepresentation, the analysis was conducted on the sole STs. Concatenated sequences recombination rates confidence intervals were calculated using likelihood curve method.
M. bolletii and M. massiliense reference strains DNA sequencing and genome assembly
The genomic DNAs of Mycobacterium abscessus subsp bolletii reference strain CIP 108541 and Mycobacterium abscessus subsp bolletii CIP 108297 (former M. massiliense reference strain) were sequenced at the Genopole of the Institut Pasteur by using the Genome Analyzer IIx (Illumina Inc., San Diego, USA) with a coverage rate of 175X and 170X, respectively. 36 bp single-end reads were generated and aligned against the reference genome of M. abscessus (EMBL accession number: CU458896)  by using MAQ . In order to prevent the presence of potential amplification contaminants, duplicated reads were removed from the alignment maps. Two reads were considered as a duplicate if they shared the same mapping position, stemmed from the same DNA strand and possessed exactly the same sequence. In the case of duplicates, the read having the best quality sum was preserved. The resulting alignment maps were then analysed by using SNIFER (https://bitbucket.org/clafooty/tango/wiki/Home) for the SNP calling, which is based on a comparison of aligned read sequences to the reference genome from mapping positions. Mismatches detected were then filtered according to 5 stringent criteria: (i) a coverage sum > 10; (ii) a substitution frequency of at least 0.89; (iii) a mean quality of mapped bases > 20 according to the Sanger format; and both mean (iv) coverage and (v) quality >20 for the 10 bases surrounding the variant (−5/+5). So as to investigate large insertion-deletion events, each short-read data set was de novo assembled using the perl script VelvetOptimiser, provided with the VELVET package .
Mycobacterium abscessus subsp. bolletii CIP 108541 genomic sequence was deposited on NCBI whole genome shotgun project with accession number JRMF00000000. Mycobacterium abscessus subsp. bolletii CIP 108297 was deposited on NCBI whole genome shotgun project with accession number JRMG00000000.
Mosaic strains DNA sequencing and genome assembly
Based on the results obtained from the Bayesian algorithm STRUCTURE, three “hybrid” strains from the MAB were selected for whole-genome sequencing (Strain M139, strain 23 and strain 137). Strain M139 came from a sputum sample from a Malaysian patient with a MAB lung infection; contigs were already generated and assembled in a former study . Strains 137 and 23 (respectively Mycobacterium abscessus subsp. bolletii 137 and Mycobacterium abscessus 23) were taken from our laboratory collection. Libraries were constructed using the Nextera Kit (Illumina) from 50 ng of DNA according to Illumina’s recommendations. Pooled libraries were sequenced on an Illumina HiSeq-2000 platform to generate 100 bp paired reads, with the TruSeq PE Cluster kit v3 and TruSeq SBS kit v3 (Illumina). All reads were pre-processed to remove low quality or artefactual nucleotides. First, all nucleotides occurring at 5′ and 3′ ends and supported by a Phred quality score < 28 were trimmed off using SICKLE (https://github.com/najoshi/sickle). Second, contaminant oligonucleotides (i.e., library adaptors) were detected and trimmed off using ALIENTRIMMER . Third, reads shorter than 45 nt after the aforementioned cleaning steps were discarded, as were those containing more than 5 % nucleotides with Phred score < 28. Finally, the program FQDUPLICATE (ftp://ftp.pasteur.fr/pub/gensoft/projects/fqtools) was used to discard every duplicate single- or paired-ends reads. A de novo assembly of the remaining reads was built with CLC Genomics Workbench version 3 (CLC Bio, Cambridge, MA). Contigs were then reordered using the MUMMER software.
The contigs of the three “hybrid” strains were compared with the three reference strains previously assembled (M. abscessus subspecies, M. massiliense subspecies, and M. bolletii subspecies) using NUCMER  and the delta-filter command (90 % minimum identity threshold on at least 400 nucleotides). The Show-Tilling script was then used to determine the order of the contigs (minimum 10 % coverage, maximal gap length 100,000 bp). The rejected contigs were manually checked and reintegrated in the final assembly if they had at least 96 % identity on > 1500 bp regions. The assembled genomes of strains M139, 23 and 137 are very similar to the reference strains (4,916,028 bp, 4,834,006 bp, and 5,011,043 bp, respectively), and represent at least 95 % of the size of the longest reference M. abscessus subsp. abscessus strain chromosome, that was sequenced by the Sanger method . Mycobacterium abscessus 23 genomic sequence was deposited on NCBI whole genome shotgun project with accession number JRMD00000000. Mycobacterium abscessus subsp. bolletii 137 was deposited on NCBI whole genome shotgun project with accession number JRME00000000.
Additionally, the 6 analysed genomes were uploaded on the Genoscope MAGE database. Core and accessory genomes were identified using a Bidirectionnal Best Hit Method, with 50 % protein identity threshold. Manual gene annotation was performed for the genetic loci of interest using global information given by this platform (Interproscan domains, SwissProt similarities, FigFam) for each gene. All regions that were identified as putative genomic islands (strain specific genetic loci of more than 8 kb) were all blasted on the NCBI NR nucleotide database, and results showing more than 80 % identity on more than 75 % of the sequence were retrieved. Graphical representation of genomic loci of interest was performed using the GENOPLOTR Package , and alignments shown were performed on genetic loci of interest using ARTEMIS software  output file.
Mobile elements detection
Mobile elements were searched in genome sequences. tRNA were identified using tRNA finder. Prophage elements were found using prophage finder (hit per prophage: 4; hit spacing: 3500). Putative insertion sequences were determined using IS finder (minimum score: 80), and manually checked with BLASTP on selected domains in order to verify the presence of recombinase or integrase genes. We checked for conjugation-associated genes in the 6 genomes (including aligned and non-aligned contigs) using traB/ftsK and virB/virD motifs on the MAGE platform (Interproscan domains, SwissProt similarities, FigFam). ESX genomic regions were identified by using a pBLAST search on all 6 reported full genomes with the M. tuberculosis ESX system core proteins EccCa (Rv3870), EccD (Rv3877) and MycP1 (Rv3883c). Type IV secretion/conjugation systems were annotated using CONJSCAN - T4SSSCAN tools .
We are grateful to Guillaume Achaz from the Atelier de Bio-informatique de Jussieu for statistical expertise; to Ghislaine Guigon from Genotyping of pathogens and Public health (Institut Pasteur) for bioinformatics expertise; to Laurence Ma for Illumina library preparation and sequencing; to Claudine Medigue and Stephane Cruveiller and David Roche from the Genoscope for the MAGE platform facilities and for technical assistance in genome scaffolding. We are grateful to the association "Vaincre la Mucoviscidose" for initial support for this work. We would also like to thank 3 anonymous reviewers for their comments that helped us to improve the manuscript.
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
- Gutierrez MC, Brisse S, Brosch R, Fabre M, Omais B, Marmiesse M, M, Supply P, Vincent V. Ancient origin and gene mosaicism of the progenitor of Mycobacterium tuberculosis. PLoS Pathog. 2005;1(1):e5.Google Scholar
- Supply P, Marceau M, Mangenot S, Roche D, Rouanet C, Khanna V, Majlessi L, Criscuolo A, Tap J, Pawlik A et al. Genomic analysis of smooth tubercle bacilli provides insights into ancestry and pathoadaptation of Mycobacterium tuberculosis. Nat Genet. 2013;45(2):172–9.View ArticlePubMedGoogle Scholar
- Rosas-Magallanes V, Deschavanne P, Quintana-Murci L, Brosch R, Gicquel B, Neyrolles O. Horizontal transfer of a virulence operon to the ancestor of Mycobacterium tuberculosis. Mol Biol Evol. 2006;23(6):1129–35.View ArticlePubMedGoogle Scholar
- Kinsella RJ, Fitzpatrick DA, Creevey CJ, McInerney JO. Fatty acid biosynthesis in Mycobacterium tuberculosis: lateral gene transfer, adaptive evolution, and gene duplication. Proc Natl Acad Sci U S A. 2003;100(18):10320–5.PubMed CentralView ArticlePubMedGoogle Scholar
- Namouchi A, Didelot X, Schock U, Gicquel B, Rocha EP. After the bottleneck: genome-wide diversification of the Mycobacterium tuberculosis complex by mutation, recombination, and natural selection. Genome Res. 2012;22(4):721–34.PubMed CentralView ArticlePubMedGoogle Scholar
- Becq J, Gutierrez MC, Rosas-Magallanes V, Rauzier J, Gicquel B, Neyrolles O, Deschavanne P. Contribution of horizontally acquired genomic islands to the evolution of the tubercle bacilli. Mol Biol Evol. 2007;24(8):1861–71.Google Scholar
- Wang J, Karnati PK, Takacs CM, Kowalski JC, Derbyshire KM. Chromosomal DNA transfer in Mycobacterium smegmatis is mechanistically different from classical Hfr chromosomal DNA transfer. Mol Microbiol. 2005;58(1):280–8.View ArticlePubMedGoogle Scholar
- Wang J, Parsons LM, Derbyshire KM. Unconventional conjugal DNA transfer in mycobacteria. Nat Genet. 2003;34(1):80–4.View ArticlePubMedGoogle Scholar
- Gray TA, Krywy JA, Harold J, Palumbo MJ, Derbyshire KM. Distributive conjugal transfer in mycobacteria generates progeny with meiotic-like genome-wide mosaicism, allowing mapping of a mating identity locus. PLoS Biol. 2013;11(7):e1001602.PubMed CentralView ArticlePubMedGoogle Scholar
- Mortimer TD, Pepperell CS. Genomic signatures of distributive conjugal transfer among mycobacteria. Genome Biol Evol. 2014;6:2489.PubMed CentralView ArticlePubMedGoogle Scholar
- Coros A, Callahan B, Battaglioli E, Derbyshire KM. The specialized secretory apparatus ESX-1 is essential for DNA transfer in Mycobacterium smegmatis. Mol Microbiol. 2008;69(4):794–808.PubMed CentralPubMedGoogle Scholar
- Simeone R, Bottai D, Brosch R. ESX/type VII secretion systems and their role in host-pathogen interaction. Curr Opin Microbiol. 2009;12(1):4–10.View ArticlePubMedGoogle Scholar
- Medjahed H, Gaillard JL, Reyrat JM. Mycobacterium abscessus: a new player in the mycobacterial field. Trends Microbiol. 2010;18(3):117–23.View ArticlePubMedGoogle Scholar
- Leung JM, Olivier KN. Nontuberculous mycobacteria: the changing epidemiology and treatment challenges in cystic fibrosis. Curr Opin Pulm Med. 2013;19(6):662–9.View ArticlePubMedGoogle Scholar
- Surette MG. The cystic fibrosis lung microbiome. Ann Am Thorac Soc. 2014;11 Suppl 1:S61–5.View ArticlePubMedGoogle Scholar
- Griffith DE. Emergence of nontuberculous mycobacteria as pathogens in cystic fibrosis. Am J Respir Crit Care Med. 2003;167(6):810–2.View ArticlePubMedGoogle Scholar
- Esther Jr CR, Esserman DA, Gilligan P, Kerr A, Noone PG. Chronic Mycobacterium abscessus infection and lung function decline in cystic fibrosis. J Cyst Fibros. 2010;9(2):117–23.View ArticlePubMedGoogle Scholar
- Roux AL, Catherinot E, Ripoll F, Soismier N, Macheras E, Ravilly S, Bellis G, Vibet MA, Le Roux E, Lemonnier L et al. Multicenter study of prevalence of nontuberculous mycobacteria in patients with cystic fibrosis in france. J Clin Microbiol. 2009;47(12):4124–8.PubMed CentralView ArticlePubMedGoogle Scholar
- Kim HS, Lee Y, Lee S, Kim YA, Sun YK. Recent trends in clinically significant nontuberculous Mycobacteria isolates at a Korean general hospital. Ann Lab Med. 2014;34(1):56–9.PubMed CentralView ArticlePubMedGoogle Scholar
- Olivier KN, Weber DJ, Wallace Jr RJ, Faiz AR, Lee JH, Zhang Y, Brown-Elliot BA, Handler A, Wilson RW, Schechter MS et al. Nontuberculous mycobacteria. I: multicenter prevalence study in cystic fibrosis. Am J Respir Crit Care Med. 2003;167(6):828–34.View ArticlePubMedGoogle Scholar
- Pierre-Audigier C, Ferroni A, Sermet-Gaudelus I, Le Bourgeois M, Offredo C, Vu-Thien H, Fauroux B, Mariani P, Munck A, Bingen E et al. Age-related prevalence and distribution of nontuberculous mycobacterial species among patients with cystic fibrosis. J Clin Microbiol. 2005;43(7):3467–70.PubMed CentralView ArticlePubMedGoogle Scholar
- Simons S, van Ingen J, Hsueh PR, Van Hung N, Dekhuijzen PN, Boeree MJ, van Soolingen D. Nontuberculous mycobacteria in respiratory tract infections, eastern Asia. Emerg Infect Dis. 2011;17(3):343–9.Google Scholar
- Huang YC, Liu MF, Shen GH, Lin CF, Kao CC, Liu PY, Shi ZY. Clinical outcome of Mycobacterium abscessus infection and antimicrobial susceptibility testing. J Microbiol Immunol Infect. 2010;43(5):401–6.Google Scholar
- Lyu J, Jang HJ, Song JW, Choi CM, Oh YM, Lee SD, Kim WS, Kim DS, Shim TS. Outcomes in patients with Mycobacterium abscessus pulmonary disease treated with long-term injectable drugs. Respir Med. 2011;105(5):781–7.Google Scholar
- Jarand J, Levin A, Zhang L, Huitt G, Mitchell JD, Daley CL. Clinical and microbiologic outcomes in patients receiving treatment for Mycobacterium abscessus pulmonary disease. Clin Infect Dis. 2011;52(5):565–71.View ArticlePubMedGoogle Scholar
- Sanguinetti M, Ardito F, Fiscarelli E, La Sorda M, D’Argenio P, Ricciotti G, Fadda G. Fatal pulmonary infection due to multidrug-resistant Mycobacterium abscessus in a patient with cystic fibrosis. J Clin Microbiol. 2001;39(2):816–9.Google Scholar
- Besada E. Rapid growing mycobacteria and TNF-alpha blockers: case report of a fatal lung infection with Mycobacterium abscessus in a patient treated with infliximab, and literature review. Clin Exp Rheumatol. 2011;29(4):705–7.PubMedGoogle Scholar
- Nessar R, Cambau E, Reyrat JM, Murray A, Gicquel B. Mycobacterium abscessus: a new antibiotic nightmare. J Antimicrob Chemother. 2012;67(4):810–8.View ArticlePubMedGoogle Scholar
- Maurer FP, Bruderer VL, Ritter C, Castelberg C, Bloemberg GV, Bottger EC. Lack of antimicrobial bactericidal activity in Mycobacterium abscessus. Antimicrob Agents Chemother. 2014;58:3828.PubMed CentralView ArticlePubMedGoogle Scholar
- Ruger K, Hampel A, Billig S, Rucker N, Suerbaum S, Bange FC. Characterization of rough and smooth morphotypes of Mycobacterium abscessus isolates from clinical specimens. J Clin Microbiol. 2014;52(1):244–50.PubMed CentralView ArticlePubMedGoogle Scholar
- Kim HY, Kim BJ, Kook Y, Yun YJ, Shin JH, Kim BJ, Kook YH. Mycobacterium massiliense is differentiated from Mycobacterium abscessus and Mycobacterium bolletii by erythromycin ribosome methyltransferase gene (erm) and clarithromycin susceptibility patterns. Microbiol Immunol. 2010;54(6):347–53.Google Scholar
- Sekizuka T, Kai M, Nakanaga K, Nakata N, Kazumi Y, Maeda S, Makino M, Hoshino Y, Kuroda M. Complete genome sequence and comparative genomic analysis of Mycobacterium massiliense JCM 15300 in the Mycobacterium abscessus group reveal a conserved genomic island MmGI-1 related to putative lipid metabolism. PLoS One. 2014;9(12):e114848.Google Scholar
- Howard ST. Recent progress towards understanding genetic variation in the Mycobacterium abscessus complex. Tuberculosis. 2013;93(Suppl):S15–20.View ArticlePubMedGoogle Scholar
- Zelazny AM, Root JM, Shea YR, Colombo RE, Shamputa IC, Stock F, Conlan S, McNulty S, Brown-Elliott BA, Wallace RJ, Jr. et al. Cohort study of molecular identification and typing of Mycobacterium abscessus, Mycobacterium massiliense, and Mycobacterium bolletii. J Clin Microbiol. 2009;47(7):1985–95.PubMed CentralView ArticlePubMedGoogle Scholar
- Kim HY, Kook Y, Yun YJ, Park CG, Lee NY, Shim TS, Kim BJ, Kook YH. Proportions of Mycobacterium massiliense and Mycobacterium bolletii strains among Korean Mycobacterium chelonae-Mycobacterium abscessus group isolates. J Clin Microbiol. 2008;46(10):3384–90.Google Scholar
- Harada T, Akiyama Y, Kurashima A, Nagai H, Tsuyuguchi K, Fujii T, Yano S, Shigeto E, Kuraoka T, Kajiki A et al. Clinical and microbiological differences between Mycobacterium abscessus and Mycobacterium massiliense lung diseases. J Clin Microbiol. 2012;50(11):3556–61.PubMed CentralView ArticlePubMedGoogle Scholar
- Tettelin H, Davidson RM, Agrawal S, Aitken ML, Shallom S, Hasan NA, Strong M, de Moura VC, De Groote MA, Duarte RS et al. High-level relatedness among Mycobacterium abscessus subsp. massiliense strains from widely separated outbreaks. Emerg Infect Dis. 2014;20(3):364–71.PubMed CentralView ArticlePubMedGoogle Scholar
- Catherinot E, Clarissou J, Etienne G, Ripoll F, Emile JF, Daffe M, Perronne C, Soudais C, Gaillard JL, Rottman M. Hypervirulence of a rough variant of the Mycobacterium abscessus type strain. Infect Immun. 2007;75(2):1055–8.Google Scholar
- Koh WJ, Jeon K, Lee NY, Kim BJ, Kook YH, Lee SH, Park YK, Kim CK, Shin SJ, Huitt GA et al. Clinical significance of differentiation of Mycobacterium massiliense from Mycobacterium abscessus. Am J Respir Crit Care Med. 2011;183(3):405–10.View ArticlePubMedGoogle Scholar
- Choo SW, Wee WY, Ngeow YF, Mitchell W, Tan JL, Wong GJ, Zhao Y, Xiao J. Genomic reconnaissance of clinical isolates of emerging human pathogen Mycobacterium abscessus reveals high evolutionary potential. Sci Rep. 2014;4:4061.Google Scholar
- Ripoll F, Pasek S, Schenowitz C, Dossat C, Barbe V, Rottman M, Macheras E, Heym B, Herrmann JL, Daffe M et al. Non mycobacterial virulence genes in the genome of the emerging pathogen Mycobacterium abscessus. PLoS One. 2009;4(6):e5660.PubMed CentralView ArticlePubMedGoogle Scholar
- Leao SC, Tortoli E, Viana-Niero C, Ueki SY, Lima KV, Lopes ML, Yubero J, Menendez MC, Garcia MJ. Characterization of mycobacteria from a major Brazilian outbreak suggests that revision of the taxonomic status of members of the Mycobacterium chelonae-M. abscessus group is needed. J Clin Microbiol. 2009;47(9):2691–8.Google Scholar
- Tan JL, Khang TF, Ngeow YF, Choo SW. A phylogenomic approach to bacterial subspecies classification: proof of concept in Mycobacterium abscessus. BMC Genomics. 2013;14:879.PubMed CentralView ArticlePubMedGoogle Scholar
- Macheras E, Roux AL, Bastian S, Leao SC, Palaci M, Sivadon-Tardy V, Gutierrez C, Richter E, Rusch-Gerdes S, Pfyffer G et al. Multilocus sequence analysis and rpoB sequencing of Mycobacterium abscessus (sensu lato) strains. J Clin Microbiol. 2011;49(2):491–9.PubMed CentralView ArticlePubMedGoogle Scholar
- Macheras E, Roux AL, Ripoll F, Sivadon-Tardy V, Gutierrez C, Gaillard JL, Heym B. Inaccuracy of single-target sequencing for discriminating species of the Mycobacterium abscessus group. J Clin Microbiol. 2009;47(8):2596–600.Google Scholar
- Sassi M, Drancourt M. Genome analysis reveals three genomospecies in Mycobacterium abscessus. BMC Genomics. 2014;15:359.PubMed CentralView ArticlePubMedGoogle Scholar
- Xia X, Xie Z, Salemi M, Chen L, Wang Y. An index of substitution saturation and its application. Mol Phylogenet Evol. 2003;26(1):1–7.View ArticlePubMedGoogle Scholar
- Bruen TC, Philippe H, Bryant D. A simple and robust statistical test for detecting the presence of recombination. Genetics. 2006;172(4):2665–81.PubMed CentralView ArticlePubMedGoogle Scholar
- McVean G, Awadalla P, Fearnhead P. A coalescent-based method for detecting and estimating recombination from gene sequences. Genetics. 2002;160(3):1231–41.PubMed CentralPubMedGoogle Scholar
- Ngeow YF, Wee WY, Wong YL, Tan JL, Ongi CS, Ng KP, Choo SW. Genomic analysis of Mycobacterium abscessus strain M139, which has an ambiguous subspecies taxonomic position. J Bacteriol. 2012;194(21):6002–3.Google Scholar
- Nash KA, Brown-Elliott BA, Wallace Jr RJ. A novel gene, erm(41), confers inducible macrolide resistance to clinical isolates of Mycobacterium abscessus but is absent from Mycobacterium chelonae. Antimicrob Agents Chemother. 2009;53(4):1367–76.PubMed CentralView ArticlePubMedGoogle Scholar
- Dumas E, Boritsch EC, Vandenbogaert M, Rodriguez de la Vega RC, Thiberge JM, Caro V, Gaillard JL, Heym B, Girard-Misguich F, Brosch R, et al. Mycobacterial pan-genome analysis suggests important role of plasmids in the radiation of type VII secretion systems. Genome Biol Evol. 2016;8(2):387–402.Google Scholar
- Leao SC, Tortoli E, Euzeby JP, Garcia MJ. Proposal that Mycobacterium massiliense and Mycobacterium bolletii be united and reclassified as Mycobacterium abscessus subsp. bolletii comb. nov., designation of Mycobacterium abscessus subsp. abscessus subsp. nov. and emended description of Mycobacterium abscessus. Int J Syst Evol Microbiol. 2011;61(Pt 9):2311–3.View ArticlePubMedGoogle Scholar
- Wirth T, Falush D, Lan R, Colles F, Mensa P, Wieler LH, Karch H, Reeves PR, Maiden MC, Ochman H et al. Sex and virulence in Escherichia coli: an evolutionary perspective. Mol Microbiol. 2006;60(5):1136–51.PubMed CentralView ArticlePubMedGoogle Scholar
- Wirth T, Morelli G, Kusecek B, van Belkum A, van der Schee C, Meyer A, Achtman M. The rise and spread of a new pathogen: Seroresistant Moraxella catarrhalis. Genome Res. 2007;17(11):1647–56.Google Scholar
- Maiden MC, Bygraves JA, Feil E, Morelli G, Russell JE, Urwin R, Zhang Q, Zhou J, Zurth K, Caugant DA et al. Multilocus sequence typing: a portable approach to the identification of clones within populations of pathogenic microorganisms. Proc Natl Acad Sci U S A. 1998;95(6):3140–5.PubMed CentralView ArticlePubMedGoogle Scholar
- Ummels R, Abdallah AM, Kuiper V, Aajoud A, Sparrius M, Naeem R, Spaink HP, van Soolingen D, Pain A, Bitter W. Identification of a novel conjugative plasmid in mycobacteria that requires both type IV and type VII secretion. mBio. 2014;5(5):e01744–14.Google Scholar
- Yang Z. PAML: a program package for phylogenetic analysis by maximum likelihood. Comput Appl Biosci. 1997;13(5):555–6.PubMedGoogle Scholar
- Abascal F, Zardoya R, Telford MJ. TranslatorX: multiple alignment of nucleotide sequences guided by amino acid translations. Nucleic Acids Res. 2010;38(Web Server issue):W7–13. doi:10.1093/nar/gkq1291. Epub 2010 Apr 1030.PubMed CentralView ArticlePubMedGoogle Scholar
- Wallace IM, O’Sullivan O, Higgins DG, Notredame C. M-Coffee: combining multiple sequence alignment methods with T-Coffee. Nucleic Acids Res. 2006;34(6):1692–9. Print 2006.PubMed CentralView ArticlePubMedGoogle Scholar
- Griffith DE, Aksamit T, Brown-Elliott BA, Catanzaro A, Daley C, Gordin F, Holland SM, Horsburgh R, Huitt G, Iademarco MF et al. An official ATS/IDSA statement: diagnosis, treatment, and prevention of nontuberculous mycobacterial diseases. Am J Respir Crit Care Med. 2007;175(4):367–416.View ArticlePubMedGoogle Scholar
- Feil EJ, Li BC, Aanensen DM, Hanage WP, Spratt BG. eBURST: inferring patterns of evolutionary descent among clusters of related bacterial genotypes from multilocus sequence typing data. J Bacteriol. 2004;186(5):1518–30.PubMed CentralView ArticlePubMedGoogle Scholar
- Xia X. DAMBE5: a comprehensive software package for data analysis in molecular biology and evolution. Mol Biol Evol. 2013;30(7):1720–8.PubMed CentralView ArticlePubMedGoogle Scholar
- Schmidt HA, Strimmer K, Vingron M, von Haeseler A. TREE-PUZZLE: maximum likelihood phylogenetic analysis using quartets and parallel computing. Bioinformatics. 2002;18(3):502–4.View ArticlePubMedGoogle Scholar
- Posada D, Crandall KA. MODELTEST: testing the model of DNA substitution. Bioinformatics. 1998;14(9):817–8.View ArticlePubMedGoogle Scholar
- Huson DH, Bryant D. Application of phylogenetic networks in evolutionary studies. Mol Biol Evol. 2006;23(2):254–67.View ArticlePubMedGoogle Scholar
- Falush D, Stephens M, Pritchard JK. Inference of population structure using multilocus genotype data: linked loci and correlated allele frequencies. Genetics. 2003;164(4):1567–87.PubMed CentralPubMedGoogle Scholar
- Falush D, Wirth T, Linz B, Pritchard JK, Stephens M, Kidd M, Blaser MJ, Graham DY, Vacher S, Perez-Perez GI et al. Traces of human migrations in Helicobacter pylori populations. Science. 2003;299(5612):1582–5.View ArticlePubMedGoogle Scholar
- Wirth T, Wang X, Linz B, Novick RP, Lum JK, Blaser M, Morelli G, Falush D, Achtman M. Distinguishing human ethnic groups by means of sequences from Helicobacter pylori: lessons from Ladakh. Proc Natl Acad Sci U S A. 2004;101(14):4746–51.Google Scholar
- Evanno G, Regnaut S, Goudet J. Detecting the number of clusters of individuals using the software STRUCTURE: a simulation study. Mol Ecol. 2005;14(8):2611–20.View ArticlePubMedGoogle Scholar
- Li H, Ruan J, Durbin R. Mapping short DNA sequencing reads and calling variants using mapping quality scores. Genome Res. 2008;18(11):1851–8.PubMed CentralView ArticlePubMedGoogle Scholar
- Zerbino DR, Birney E. Velvet: algorithms for de novo short read assembly using de Bruijn graphs. Genome Res. 2008;18(5):821–9.PubMed CentralView ArticlePubMedGoogle Scholar
- Criscuolo A, Brisse S. AlienTrimmer: a tool to quickly and accurately trim off multiple short contaminant sequences from high-throughput sequencing reads. Genomics. 2013;102(5-6):500–6.View ArticlePubMedGoogle Scholar
- Kurtz S, Phillippy A, Delcher AL, Smoot M, Shumway M, Antonescu C, Salzberg SL. Versatile and open software for comparing large genomes. Genome Biol. 2004;5(2):R12.Google Scholar
- Guy L, Kultima JR, Andersson SG. genoPlotR: comparative gene and genome visualization in R. Bioinformatics. 2010;26(18):2334–5.PubMed CentralView ArticlePubMedGoogle Scholar
- Carver T, Harris SR, Berriman M, Parkhill J, McQuillan JA. Artemis: an integrated platform for visualization and analysis of high-throughput sequence-based experimental data. Bioinformatics. 2012;28(4):464–9.PubMed CentralView ArticlePubMedGoogle Scholar
- Guglielmini J, Quintais L, Garcillan-Barcia MP, de la Cruz F, Rocha EP. The repertoire of ICE in prokaryotes underscores the unity, diversity, and ubiquity of conjugation. PLoS Genet. 2011;7(8):e1002222.PubMed CentralView ArticlePubMedGoogle Scholar
- Rosenberg NA. Distruct: a program for the graphical display of population structure. Mol Ecol Notes. 2004;4:137–8.View ArticleGoogle Scholar
- Krzywinski M, Schein J, Birol I, Connors J, Gascoyne R, Horsman D, Jones SJ, Marra MA. Circos: an information aesthetic for comparative genomics. Genome Res. 2009;19(9):1639–45.Google Scholar