Genomics and Machine Learning for Taxonomy Consensus: The Mycobacterium tuberculosis Complex Paradigm

The shortcut of mycobacterial interspersed repetitive unit-variable number tandem repeat typing for Mycobacterium tuberculosis differentiation

The 24-loci mycobacterial interspersed repetitive unit-variable number tandem repeat (MIRU-VNTR) genotyping has been used as an international standard method for Mycobacterium tuberculosis (Mtb) genotyping. However, different optimized VNTR loci sets for improving the discrimination of specific Mtb genotypes have been proposed. In this regard, we investigated the efficacy of accumulation of the percentage differences (APDs) compared with the least absolute shrinkage and selection operator (LASSO) regression strategy to identify a customized genotype-specific VNTR loci set which provides a resolution comparable to 24-loci MIRU-VNTR in divergent Mtb populations. We utilized Spoligotyping and 24-loci MIRU-VNTR typing for genotyping 306 Mtb isolates. The APD and LASSO regression approaches were used to identify a customized VNTR set in our studied isolates. Besides, the Hunter-Gaston discriminatory index (HGDI), sensitivity, and specificity of each selected loci set were calculated based on both strategies. The selected loci based on LASSO regression compared with APD-based loci showed a better discriminatory power for identifying all studied genotypes except for T genotype, which APD-based loci showed promising discriminative power. Our findings suggested the LASSO regression rather than the APD approach is more effective in the determination of possible discriminative VNTR loci set to precise discrimination of our studied Mtb population and may be beneficial to be used in finding reduced number loci sets in other Mtb genotypes or sublineages. Moreover, we proposed customized genotype-specific MIRU-VNTR loci sets based on the LASSO regression and APD approaches for precise Mtb strains identification. As the proposed VNTR sets offered a comparable discriminatory power to the standard 24 MIRU-VNTR loci set could be promising alternatives to the standard genotyping for using in resource-limited settings.

Molecular epidemiology of Mycobacterium tuberculosis in Brazil before the whole genome sequencing era: a literature review

Molecular-typing can help in unraveling epidemiological scenarios and improvement for disease control strategies. A literature review of Mycobacterium tuberculosis transmission in Brazil through genotyping on 56 studies published from 1996-2019 was performed. The clustering rate for mycobacterial interspersed repetitive units - variable tandem repeats (MIRU-VNTR) of 1,613 isolates were: 73%, 33% and 28% based on 12, 15 and 24-loci, respectively; while for RFLP-IS6110 were: 84% among prison population in Rio de Janeiro, 69% among multidrug-resistant isolates in Rio Grande do Sul, and 56.2% in general population in São Paulo. These findings could improve tuberculosis (TB) surveillance and set up a solid basis to build a database of Mycobacterium genomes.

CRISPRbuilder-TB: "CRISPR-builder for tuberculosis". Exhaustive reconstruction of the CRISPR locus in mycobacterium tuberculosis complex using SRA

Mycobacterium tuberculosis complex (MTC) CRISPR locus diversity has long been studied solely investigating the presence/absence of a known set of spacers. Unveiling the genetic mechanisms of its evolution requires a more exhaustive reconstruction in a large amount of representative strains.

In this article, we point out and resolve, with a new pipeline, the problem of CRISPR reconstruction based directly on short read sequences in M. tuberculosis. We first show that the process we set up, that we coin as “CRISPRbuilder-TB” (https://github.com/cguyeux/CRISPRbuilder-TB), allows an efficient reconstruction of simulated or real CRISPRs, even when including complex evolutionary steps like the insertions of mobile elements. Compared to more generalist tools, the whole process is much more precise and robust, and requires only minimal manual investigation. Second, we show that more than 1/3 of the currently complete genomes available for this complex in the public databases contain largely erroneous CRISPR loci. Third, we highlight how both the classical experimental in vitro approach and the basic in silico spoligotyping provided by existing analytic tools miss a whole diversity of this locus in MTC, by not capturing duplications, spacer and direct repeats variants, and IS6110 insertion locations. This description is extended in a second article that describes MTC-CRISPR diversity and suggests general rules for its evolution. This work opens perspectives for an in-depth exploration of M. tuberculosis CRISPR loci diversity and of mechanisms involved in its evolution and its functionality, as well as its adaptation to other CRISPR locus-harboring bacterial species.

In this article, we tackle the bioinformatical issue of the reconstruction of the Mycobacterium tuberculosis complex CRISPR locus using short read sequences without requiring genome assembly. We first show that many complete genomes, as found in public databases and often reconstructed by de novo assemblies, often contain errors on this locus as well as on other repeated sequences. We provide an in-depth description of our new method, designated as ‘CRISPRbuilder-TB’, and we show that our method provides much more exhaustive and reliable information (on DR variants, spacer diversity, global structure) than Crass and CRISPR_detector. The new and unsuspected genomic diversity we detected is described in a companion paper. Scripts are available to adapt the tool to other species.

Novel methods included in SpolLineages tool for fast and precise prediction of Mycobacterium tuberculosis complex spoligotype families

Bioinformatic tools are currently being developed to better understand the Mycobacterium tuberculosis complex (MTBC). Several approaches already exist for the identification of MTBC lineages using classical genotyping methods such as mycobacterial interspersed repetitive units-variable number of tandem DNA repeats and spoligotyping-based families. In the recently released SITVIT2 proprietary database of the Institut Pasteur de la Guadeloupe, a large number of spoligotype families were assigned by either manual curation/expertise or using an in-house algorithm. In this study, we present two complementary data-driven approaches allowing fast and precise family prediction from spoligotyping patterns. The first one is based on data transformation and the use of decision tree classifiers. In contrast, the second one searches for a set of simple rules using binary masks through a specifically designed evolutionary algorithm. The comparison with the three main approaches in the field highlighted the good performances of our contributions and the significant runtime gain. Finally, we propose the 'SpolLineages' software tool (https://github.com/dcouvin/SpolLineages), which implements these approaches for MTBC spoligotype families' identification.

A guide to machine learning for bacterial host attribution using genome sequence data

With the ever-expanding number of available sequences from bacterial genomes, and the expectation that this data type will be the primary one generated from both diagnostic and research laboratories for the foreseeable future, then there is both an opportunity and a need to evaluate how effectively computational approaches can be used within bacterial genomics to predict and understand complex phenotypes, such as pathogenic potential and host source. This article applied various quantitative methods such as diversity indexes, pangenome-wide association studies (GWAS) and dimensionality reduction techniques to better understand the data and then compared how well unsupervised and supervised machine learning (ML) methods could predict the source host of the isolates. The study uses the example of the pangenomes of 1203 Salmonella enterica serovar Typhimurium isolates in order to predict 'host of isolation' using these different methods. The article is aimed as a review of recent applications of ML in infection biology, but also, by working through this specific dataset, it allows discussion of the advantages and drawbacks of the different techniques. As with all such sub-population studies, the biological relevance will be dependent on the quality and diversity of the input data. Given this major caveat, we show that supervised ML has the potential to add real value to interpretation of bacterial genomic data, as it can provide probabilistic outcomes for important phenotypes, something that is very difficult to achieve with the other methods.

Prediction of the hidden genotype of mixed infection strains in Iranian tuberculosis patients

BACKGROUND

Patients with mixed-strain Mycobacterium tuberculosis infections may be at a high risk of poor treatment outcomes. However, the mechanisms through which mixed infections affect the clinical manifestations are not well recognized. Evidence suggests that failure to detect the pathogen diversity within the host can influence the clinical results. We aimed to investigate the effects of different genotypes in mixed infections and determine their relationship with heteroresistance in the treatment of Iranian tuberculosis patients.

METHODS

One of the genotypes was identified in the culture and another genotype pattern in the mixed infection was predicted by comparing the pattern of MIRU-VNTR between the clinical specimens and their respective cultures in each patient. For all patients, the drug susceptibility testing was carried out on three single colonies from each clinical sample. The follow-up of patients was carried out during six months of treatment.

RESULTS

Based on MIRU-VNTR profiles of clinical samples, we showed that 55.6% (25/45) of the Iranian patients included in the study had mixed infections. Patients with mixed infections had a higher rate of treatment failure, compared to others (P=0.03). By comparing clinical sample profiles to profiles obtained after culture, we were able to distinguish between major and hidden genotypes. Among hidden genotypes, Haarlem (L4.1.2) and Beijing (L2) were associated to treatment failure (6/8 patients).

CONCLUSIONS

To conclude, we propose a procedure using the MIRU-VNTR method to identify the different genotypes in mixed infections. The present findings suggest that genotypes with potentially higher pathogenicity may not be detected when performing experimental culture in patients with mixed infections.

Enhancing the differentiation of specific genotypes in Mycobacterium tuberculosis population

Today, significant attention is directed towards the global lineages and sublineages of Mycobacterium tuberculosis (Mtb). NEW-1 (SIT 127) and CAS1-Delhi (SIT 26) strains are recognized as growing and circulating Mtb genotypes, especially in Asian countries. It is crucial to develop or enhance Mtb genotyping methods for a more accurate and simple differentiation of these families. We used 24-loci mycobacterial interspersed repetitive unit-variable number tandem repeat (MIRU-VNTR) typing for genotyping 217 Mtb isolates. To select the optimal MIRU-VNTR loci, we calculated the Hunter-Gaston discriminatory index (HGDI), allelic diversity, and accumulation of percentage differences (APDs) between the strains among different groups of genotypes (NEW-1 and non-NEW-1; CAS1-Delhi and non-CAS). Finally, the minimum spanning tree was constructed for clustering analysis. In the NEW-1 population, loci with APD > 60% were found to have a high discriminatory power. VNTR loci with APD > 50% showed high discrimination power for the CAS population. Our findings suggest that APDs, which are valuable for the selection of VNTR loci sets, may improve the discriminatory power of MIRU-VNTR typing for identification of Mtb genotypes in specific regions.

Diversity of Mycobacterium tuberculosis in the Middle Fly District of Western Province, Papua New Guinea: microbead-based spoligotyping using DNA from Ziehl-Neelsen-stained microscopy preparations

Tuberculosis remains the world's leading cause of death from an infectious agent, and is a serious health problem in Papua New Guinea (PNG) with an estimated 36,000 new cases each year. This study describes the genetic diversity of Mycobacterium tuberculosis among tuberculosis patients in the Balimo/Bamu region in the Middle Fly District of Western Province in PNG, and investigates rifampicin resistance-associated mutations. Archived Ziehl-Neelsen-stained sputum smears were used to conduct microbead-based spoligotyping and assess genotypic resistance. Among the 162 samples included, 80 (49.4%) generated spoligotyping patterns (n = 23), belonging predominantly to the L2 Lineage (44%) and the L4 Lineage (30%). This is consistent with what has been found in other PNG regions geographically distant from Middle Fly District of Western Province, but is different from neighbouring South-East Asian countries. Rifampicin resistance was identified in 7.8% of the successfully sequenced samples, with all resistant samples belonging to the L2/Beijing Lineage. A high prevalence of mixed L2/L4 profiles was suggestive of polyclonal infection in the region, although this would need to be confirmed. The method described here could be a game-changer in resource-limited countries where large numbers of archived smear slides could be used for retrospective (and prospective) studies of M. tuberculosis genetic epidemiology.

Towards better prediction of Mycobacterium tuberculosis lineages from MIRU-VNTR data

The determination of lineages from strain-based molecular genotyping information is an important problem in tuberculosis. Mycobacterial interspersed repetitive unit-variable number tandem repeat (MIRU-VNTR) typing is a commonly used molecular genotyping approach that uses counts of the number of times pre-specified loci repeat in a strain. There are three main approaches for determining lineage based on MIRU-VNTR data - one based on a direct comparison to the strains in a curated database, and two others, on machine learning algorithms trained on a large collection of labeled data. All existing methods have limitations. The direct approach imposes an arbitrary threshold on how much a database strain can differ from a given one to be informative. On the other hand, the machine learning-based approaches require a substantial amount of labeled data. Notably, all three methods exhibit suboptimal classification accuracy without additional data. We explore several computational approaches to address these limitations. First, we show that eliminating the arbitrary threshold improves the performance of the direct approach. Second, we introduce RuleTB, an alternative direct method that proposes a concise set of rules for determining lineages. Lastly, we propose StackTB, a machine learning approach that requires only a fraction of the training data to outperform the accuracy of both existing machine learning methods. Our approaches demonstrate superior performance on a training dataset collected in New York City over 10 years, and the improvement in performance translates to a held-out testing set. We conclude that our methods provide opportunities for improving the determination of pathogenic lineages based on MIRU-VNTR data.

Mycobacterium tuberculosis lineage 1 genetic diversity in Pará, Brazil, suggests common ancestry with east-African isolates potentially linked to historical slave trade

Lineage 1 (L1) is one of seven Mycobacterium tuberculosis complex (MTBC) lineages. The objective of this study was to improve the complex taxonomy of L1 using phylogenetic SNPs, and to look for the origin of the main L1 sublineage prevalent in Para, Brazil. We developed a high-throughput SNPs-typing assay based on 12-L1-specific SNPs. This assay allowed us to experimentally retrieve SNP patterns on nine of these twelve SNPs in 277 isolates previously tentatively assigned to L1 spoligotyping-based sublineages. Three collections were used: Pará-Brazil (71); RIVM, the Netherlands (102), Madagascar (104). One-hundred more results were generated in Silico using the PolyTB database. Based on the final SNPs combination, the samples were classified into 11 clusters (C1-C11). Most isolates within a SNP-based cluster shared a mutual spoligotyping-defined lineage. However, L1/EAI1-SOM (SIT48) and L1/EAI6-BGD1 (SIT591) showed a poor correlation with SNP data and are not monophyletic. L1/EAI8-MDG and L1/EAI3-IND belonged to C5; this result suggests that they share a common ancestor. L1.1.3/SIT129, a spoligotype pattern found in SNPs-cluster C6, was found to be shared between Pará/Brazil and Malawi. SIT129 was independently found to be highly prevalent in Mozambique, which suggests a migration history from East-Africa to Brazil during the 16th-18th slave trade period to Northern Brazil.

Universal genotyping reveals province-level differences in the molecular epidemiology of tuberculosis

OBJECTIVES

Compare the molecular epidemiology of tuberculosis (TB) between two large Canadian provinces-Ontario and British Columbia (BC)-to identify genotypic clusters within and across both provinces, allowing for an improved understanding of genotype data and providing context to more accurately identify clusters representing local transmission.

DESIGN

We compared 24-locus Mycobacterial Interspersed Repetitive Units-Variable Number of Tandem Repeats (MIRU-VNTR) genotyping for 3,314 Ontario and 1,602 BC clinical Mycobacterium tuberculosis isolates collected from 2008 through 2014. Laboratory data for each isolate was linked to case-level records to obtain clinical and demographic data.

RESULTS

The demographic characteristics of persons with TB varied between provinces, most notably in the proportion of persons born outside Canada, which was reflected in the large number of unique genotypes (n = 3,461). The proportion of clustered isolates was significantly higher in BC. Substantial clustering amongst non-Lineage 4 TB strains was observed within and across the provinces. Only two large clusters (≥10 cases/cluster) representing within province transmission had interprovincial genotype matches.

CONCLUSION

We recommend expanding analysis of shared genotypes to include neighbouring jurisdictions, and implementing whole genome sequencing to improve identification of TB transmission, recognize outbreaks, and monitor changing trends in TB epidemiology.

Simplified Model to Survey Tuberculosis Transmission in Countries Without Systematic Molecular Epidemiology Programs

Systematic molecular/genomic epidemiology studies for tuberculosis surveillance cannot be implemented in many countries. We selected Panama as a model for an alternative strategy. Mycobacterial interspersed repetitive unit-variable-number tandem-repeat (MIRU-VNTR) analysis revealed a high proportion (50%) of Mycobacterium tuberculosis isolates included in 6 clusters (A-F) in 2 provinces (Panama and Colon). Cluster A corresponded to the Beijing sublineage. Whole-genome sequencing (WGS) differentiated clusters due to active recent transmission, with low single-nucleotide polymorphism-based diversity (cluster C), from clusters involving long-term prevalent strains with higher diversity (clusters A, B). Prospective application in Panama of 3 tailored strain-specific PCRs targeting marker single-nucleotide polymorphisms identified from WGS data revealed that 31.4% of incident cases involved strains A-C and that the Beijing strain was highly represented and restricted mainly to Colon. Rational integration of MIRU-VNTR, WGS, and tailored strain-specific PCRs could be a new model for tuberculosis surveillance in countries without molecular/genomic epidemiology programs.

Genetic characterization of Mycobacterium tuberculosis complex isolates circulating in Abuja, Nigeria

Objective

Nigeria ranks fourth among the high tuberculosis (TB) burden countries. This study describes the prevalence of drug resistance and the genetic diversity of Mycobacterium tuberculosis in Abuja’s Federal Capital Territory.

Materials and methods

Two hundred and seventy-eight consecutive sputum samples were collected from adults with presumptive TB during 2013–2014. DNA was extracted from Löwenstein–Jensen cultures and analyzed for the identification of nontuberculous mycobacteria species, detection of drug resistance with line probe assays, and high-throughput spacer oligonucleotide typing (spoligotyping) using microbead-based hybridization.

Results

Two hundred and two cultures were positive for M. tuberculosis complex, 24 negative, 38 contaminated, and 15 positive for nontuberculous mycobacteria. Five (2.5%) M. tuberculosis complex isolates were resistant to rifampicin (RIF) and isoniazid (multidrug resistant), nine (4.5%) to RIF alone, and 15 (7.4%) to isoniazid alone; two RIF-resistant isolates were also resistant to fluoroquinolones and ethambutol, and one multidrug resistant isolate was also resistant to ethambutol. Among the 180 isolates with spoligotyping results, 164 (91.1%) were classified as lineage 4 (Euro-American), 13 (7.2%) as lineage 5 (West African 1), two (1.1%) as lineage 2 (East Asia), and one (0.6%) as lineage 6 (West African 2). One hundred and fifty-six (86.7%) isolates were grouped in 17 clusters (2–108 isolates/cluster), of which 108 (60.0%) were grouped as L4.6.2/Cameroon (spoligotype international type 61).

Conclusion

The description of drug resistance prevalence and genetic diversity of M. tuberculosis in this study may be useful for improving TB control in Nigeria.

Genetic diversity of Mycobacterium tuberculosis complex isolates circulating in an area with high tuberculosis incidence: Using 24-locus MIRU-VNTR method

We aimed to determine the genetic diversity, phylogenetic relationship and transmission dynamics of Mycobacterium tuberculosis complex (MTBC) genotypes in an area with high tuberculosis (TB) incidence. A set of 164 MTBC isolates from new TB patients of Golestan province, Iran, were subjected to genotyping using the standard 24-locus MIRU-VNTR method. Recent TB transmission was evaluated and phylogenetic relationships were analysed by minimum spanning tree and cluster-graph methods. Among the 164 isolates, 132 distinct patterns were detected. The 48 clustered isolates (29.3%) were distributed into 16 clusters ranging in size from 2 to 12 isolates. The most frequent genotype was Central Asian Strain/Delhi (CAS/Delhi) (n = 67, 40.8%), followed by NEW-1 (n = 53, 32.3%) and Beijing (n = 19, 11.6%) genotypes. Thirty five (72.9%) of NEW-1 isolates were recovered from immigrant patients and 84.2% (n = 16) of Beijing genotypes recovered from native cases. Statistically significant association was found between clustering and smoking (p = 0.047), drug addiction (p = 0.01) and prison history (p = 0.003). The estimated proportion of recent transmission was 19.5%. Presence of highly diverse MTBC isolates circulating in this province without a dominant genotype might be a consequence of importation of various genotypes in this area.

Systematic evaluation of supervised classifiers for fecal microbiota-based prediction of colorectal cancer

Predicting colorectal cancer (CRC) based on fecal microbiota presents a promising method for non-invasive screening of CRC, but the optimization of classification models remains an unaddressed question. The purpose of this study was to systematically evaluate the effectiveness of different supervised machine-learning models in predicting CRC in two independent eastern and western populations. The structures of intestinal microflora in feces in Chinese population (N = 141) were determined by 454 FLX pyrosequencing, and different supervised classifiers were employed to predict CRC based on fecal microbiota operational taxonomic unit (OTUs). As a result, Bayes Net and Random Forest displayed higher accuracies than other algorithms in both populations, although Bayes Net was found with a lower false negative rate than that of Random Forest. Gut microbiota-based prediction was more accurate than the standard fecal occult blood test (FOBT), and the combination of both approaches further improved the prediction accuracy. Moreover, when unclassified OTUs were used as input, the BayesDMNB text algorithm achieved higher accuracy in the Chinese population (AUC=0.994). Taken together, our results suggest that Bayes Net classification model combined with unclassified OTUs may present an accurate method for predicting CRC based on the compositions of gut microbiota.

Optimizing and accelerating the assignation of lineages in Mycobacterium tuberculosis using novel alternative single-tube assays

The assignation of lineages in Mycobacterium tuberculosis (MTB) provides valuable information for evolutionary and phylogeographic studies and makes for more accurate knowledge of the distribution of this pathogen worldwide. Differences in virulence have also been found for certain lineages. MTB isolates were initially assigned to lineages based on data obtained from genotyping techniques, such as spoligotyping or MIRU-VNTR analysis, some of which are more suitable for molecular epidemiology studies. However, since these methods are subject to a certain degree of homoplasy, other criteria have been chosen to assign lineages. These are based on targeting robust and specific SNPs for each lineage. Here, we propose two newly designed multiplex targeting methods—both of which are single-tube tests—to optimize the assignation of the six main lineages in MTB. The first method is based on ASO-PCR and offers an inexpensive and easy-to-implement assay for laboratories with limited resources. The other, which is based on SNaPshot, enables more refined standardized assignation of lineages for laboratories with better resources. Both methods performed well when assigning lineages from cultured isolates from a control panel, a test panel, and a problem panel from an unrelated population, Mexico, which included isolates in which standard genotyping was not able to classify lineages. Both tests were also able to assign lineages from stored isolates, without the need for subculture or purification of DNA, and even directly from clinical specimens with a medium-high bacilli burden. Our assays could broaden the contexts where information on lineages can be acquired, thus enabling us to quickly update data from retrospective collections and to merge data with those obtained at the time of diagnosis of a new TB case.

Evaluation of spoligotyping, SNPs and customised MIRU-VNTR combination for genotyping Mycobacterium tuberculosis clinical isolates in Madagascar

BACKGROUND

Combining different molecular typing methods for Mycobacterium tuberculosis complex (MTBC) can be a powerful tool for molecular epidemiology-based investigation of TB. However, the current standard method that provides high discriminatory power for such a combination, mycobacterial interspersed repetitive units-variable numbers of tandem repeats typing (MIRU-VNTR), is laborious, time-consuming and often too costly for many resource-limited laboratories. We aimed to evaluate a reduced set of loci for MIRU-VNTR typing in combination with spoligotyping and SNP-typing for routine molecular epidemiology of TB.

METHOD

Spoligotyping and SNP-typing, in combination with the 15 loci MIRU-VNTR typing, were first used to type clinical MTBC isolates (n = 158) from Madagascar. A step by step reduction of MIRU-VNTR loci number was then performed according to the Hunter and Gaston Discriminatory Index (HGDI) and to the Principal component analysis (PCA) correlation with the spoligotype profiles to evaluate the discrimination power inside the generated spoligotype clusters. The 15 MIRU-VNTR was used as reference and SNP-typing was used to determine the main MTBC lineages.

RESULTS

Of the 158 clinical isolates studied, the SNP-typing classified 23 into Lineage 1 (14.6%), 31 into Lineage 2 (19.6%), 23 into Lineage 3 (14.6%) and 81 into Lineage 4 strains (51.3%). 37 different spoligotypes profiles were obtained, 15 of which were unique and 20 in clusters. 15-loci MIRU-VNTR typing revealed 144 different genotypes: 132 isolates had a unique MIRU-VNTR profile and 27 isolates were grouped into 12 clusters. After a stepwise reduction of the MIRU-VNTR loci number within each main spoligotype families, three different sets composed of 5 customised MIRU-VNTR loci had a similar discrimination level to the reference 15 loci MIRU-VNTR in lineage 1, lineage 2 and lineage 3. For lineage 4, a set of 4 and 3 MIRU-VNTR loci were proposed to subtype the Harleem and LAM spoligotype families, respectively. For the T spoligotype family, a set of 5 MIRU-VNTR loci was proposed.

CONCLUSION

According to the lineages and the spoligotype families, the number of MIRU-VNTR loci can be reduced to get an optimal classification of MTBC. These customized sets of MIRU-VNTR loci reduce workload and save resources while maintaining optimal discriminatory power.

Reinstating Mycobacterium massiliense and Mycobacterium bolletii as species of the Mycobacterium abscessus complex

TheMycobacterium abscessus complex is a group of rapidly growing, multiresistant mycobacteria previously divided into three species. Proposal for the union of Mycobacterium bolletii and Mycobacterium massiliense into one subspecies, so-called M. abscessus subsp. massiliense, created much confusion about the routine identification and reporting of M. abscessus clinical isolates for clinicians. Results derived from multigene sequencing unambiguously supported the reinstatement of M. massiliense and M. bolletii as species, culminating in the presence of erm(41)-encoded macrolide resistance in M. bolletii. Present genome-based analysis unambiguously supports the reinstatement of M. massiliense and M. bolletii as species after the average nucleotide identity values of 96.7 % for M. abscessus versus M. bolletii, and 96.4 % for M. abscessus versus M. massiliense, and the 96.6 % identity between M. bolletii and M. massiliense was put into the perspective of a larger, 28-species analysis. Accordingly, DNA-DNA hybridization values predicted by the complete rpoB gene sequencing analysis were between 68.7 and 72.3 % in this complex. These genomic data as well as the phenotypic characteristics prompted us to propose to reinstate the previously known M. massiliense and M. bolletii into two distinct species among the M. abscessus complex.

Turkish and Japanese Mycobacterium tuberculosis sublineages share a remote common ancestor

Two geographically distant M. tuberculosis sublineages, Tur from Turkey and T3-Osaka from Japan, exhibit partially identical genotypic signatures (identical 12-loci MIRU-VNTR profiles, distinct spoligotyping patterns). We investigated T3-Osaka and Tur sublineages characteristics and potential genetic relatedness, first using MIRU-VNTR locus analysis on 21 and 25 samples of each sublineage respectively, and second comparing Whole Genome Sequences of 8 new samples to public data from 45 samples uncovering human tuberculosis diversity. We then tried to date their Most Recent Common Ancestor (MRCA) using three calibrations of SNP accumulation rate (long-term=0.03SNP/genome/year, derived from a tuberculosis ancestor of around 70,000years old; intermediate=0.2SNP/genome/year derived from a Peruvian mummy; short-term=0.5SNP/genome/year). To disentangle between these scenarios, we confronted the corresponding divergence times with major human history events and knowledge on human genetic divergence. We identified relatively high intrasublineage diversity for both T3-Osaka and Tur. We definitively proved their monophyly; the corresponding super-sublineage (referred to as "T3-Osa-Tur") shares a common ancestor with T3-Ethiopia and Ural sublineages but is only remotely related to other Euro-American sublineages such as X, LAM, Haarlem and S. The evolutionary scenario based on long-term evolution rate being valid until T3-Osa-Tur MRCA was not supported by Japanese fossil data. The evolutionary scenario relying on short-term evolution rate since T3-Osa-Tur MRCA was contradicted by human history and potential traces of past epidemics. T3-Osaka and Tur sublineages were found likely to have diverged between 800y and 2000years ago, potentially at the time of Mongol Empire. Altogether, this study definitively proves a strong genetic link between Turkish and Japanese tuberculosis. It provides a first hypothesis for calibrating TB Euro-American lineage molecular clock; additional studies are needed to reliably date events corresponding to intermediate depths in tuberculosis phylogeny.