Mycobacterium tuberculosis resistance prediction and lineage classification from genome sequencing: comparison of automated analysis tools

A Profile of Drug-Resistant Mutations in Mycobacterium tuberculosis Isolates from Guangdong Province, China

Guangdong Province, China's largest economy, has a high incidence of tuberculosis (TB). At present, there are few reports on the distribution, transmission and drug resistance of Mycobacterium tuberculosis (Mtb) strains in this region. In this study, we performed minimum inhibitory concentration testing for 14 anti-TB drugs and whole-genome sequencing of 713 clinical Mtb isolates from 20,662 sputum culture-positive tuberculosis patients registered at 31 tuberculosis drug resistance surveillance sites covering 20 cities in Guangdong Province from 2016 to 2018. Moreover, we evaluated genome-wide associations between mutations and drug resistance, and further investigated the differences in the MICs of mutations. The epidemiology, drug-resistant phenotypes and whole genome sequencing data of 713 clinical Mtb isolates were analyzed, revealing the lineage distribution and drug-resistant gene profiles in Guangdong Province. WGS combined with quantitative MIC measurements identified several novel loci associated with resistance, of which 16 loci were found to be related to resistance to more than one drug. This study analyzed the lineage distribution, prevalence characteristics and resistance-corresponding gene profiles of Mtb isolates in Guangdong province, and provided a theoretical basis for the formulation of tuberculosis prevention and control policy in the province.

Supplementary Information

The online version contains supplementary material available at 10.1007/s12088-024-01236-3.

TB-DROP: deep learning-based drug resistance prediction of Mycobacterium tuberculosis utilizing whole genome mutations

The most widely practiced strategy for constructing the deep learning (DL) prediction model for drug resistance of Mycobacterium tuberculosis (MTB) involves the adoption of ready-made and state-of-the-art architectures usually proposed for non-biological problems. However, the ultimate goal is to construct a customized model for predicting the drug resistance of MTB and eventually for the biological phenotypes based on genotypes. Here, we constructed a DL training framework to standardize and modularize each step during the training process using the latest tensorflow 2 API. A systematic and comprehensive evaluation of each module in the three currently representative models, including Convolutional Neural Network, Denoising Autoencoder, and Wide & Deep, which were adopted by CNNGWP, DeepAMR, and WDNN, respectively, was performed in this framework regarding module contributions in order to assemble a novel model with proper dedicated modules. Based on the whole-genome level mutations, a de novo learning method was developed to overcome the intrinsic limitations of previous models that rely on known drug resistance-associated loci. A customized DL model with the multilayer perceptron architecture was constructed and achieved a competitive performance (the mean sensitivity and specificity were 0.90 and 0.87, respectively) compared to previous ones. The new model developed was applied in an end-to-end user-friendly graphical tool named TB-DROP (TuBerculosis Drug Resistance Optimal Prediction: https://github.com/nottwy/TB-DROP ), in which users only provide sequencing data and TB-DROP will complete analysis within several minutes for one sample. Our study contributes to both a new strategy of model construction and clinical application of deep learning-based drug-resistance prediction methods.

Use of Whole Genome Sequencing for Mycobacterium tuberculosis Complex Antimicrobial Susceptibility Testing: From Sequence Data to Resistance Profiles

Whole genome sequencing of Mycobacterium tuberculosis complex (MTBC) isolates has been shown to provide accurate predictions for resistance and susceptibility for many first- and second-line anti-tuberculosis drugs. However, bioinformatic pipelines and mutation catalogs to predict antimicrobial resistances in MTBC isolates are often customized and detailed protocols are difficult to access. Here, we provide a step-by-step workflow for the processing and interpretation of short-read sequencing data and give an overview of available analysis pipelines.

Evolution and transmission of antibiotic resistance is driven by Beijing lineage Mycobacterium tuberculosis in Vietnam

IMPORTANCE

Drug-resistant tuberculosis (TB) infection is a growing and potent concern, and combating it will be necessary to achieve the WHO's goal of a 95% reduction in TB deaths by 2035. While prior studies have explored the evolution and spread of drug resistance, we still lack a clear understanding of the fitness costs (if any) imposed by resistance-conferring mutations and the role that Mtb genetic lineage plays in determining the likelihood of resistance evolution. This study offers insight into these questions by assessing the dynamics of resistance evolution in a high-burden Southeast Asian setting with a diverse lineage composition. It demonstrates that there are clear lineage-specific differences in the dynamics of resistance acquisition and transmission and shows that different lineages evolve resistance via characteristic mutational pathways.

Direct detection of drug-resistant Mycobacterium tuberculosis using targeted next generation sequencing

Mycobacterium tuberculosis complex (MTBC) infections are treated with combinations of antibiotics; however, these regimens are not as efficacious against multidrug and extensively drug resistant MTBC. Phenotypic (growth-based) drug susceptibility testing on slow growing bacteria like MTBC requires many weeks to months to complete, whereas sequencing-based approaches can predict drug resistance (DR) with reduced turnaround time. We sought to develop a multiplexed, targeted next generation sequencing (tNGS) assay that can predict DR and can be performed directly on clinical respiratory specimens. A multiplex PCR was designed to amplify a group of thirteen full-length genes and promoter regions with mutations known to be involved in resistance to first- and second-line MTBC drugs. Long-read amplicon libraries were sequenced with Oxford Nanopore Technologies platforms and high-confidence resistance mutations were identified in real-time using an in-house developed bioinformatics pipeline. Sensitivity, specificity, reproducibility, and accuracy of the tNGS assay was assessed as part of a clinical validation study. In total, tNGS was performed on 72 primary specimens and 55 MTBC-positive cultures and results were compared to clinical whole genome sequencing (WGS) performed on paired patient cultures. Complete or partial susceptibility profiles were generated from 82% of smear positive primary specimens and the resistance mutations identified by tNGS were 100% concordant with WGS. In addition to performing tNGS on primary clinical samples, this assay can be used to sequence MTBC cultures mixed with other mycobacterial species that would not yield WGS results. The assay can be effectively implemented in a clinical/diagnostic laboratory with a two to three day turnaround time and, even if batched weekly, tNGS results are available on average 15 days earlier than culture-derived WGS results. This study demonstrates that tNGS can reliably predict MTBC drug resistance directly from clinical specimens or cultures and provide critical information in a timely manner for the appropriate treatment of patients with DR tuberculosis.

Bioinformatics toolbox for exploring target mutation-induced drug resistance

Drug resistance is increasingly among the main issues affecting human health and threatening agriculture and food security. In particular, developing approaches to overcome target mutation-induced drug resistance has long been an essential part of biological research. During the past decade, many bioinformatics tools have been developed to explore this type of drug resistance, and they have become popular for elucidating drug resistance mechanisms in a low cost, fast and effective way. However, these resources are scattered and underutilized, and their strengths and limitations have not been systematically analyzed and compared. Here, we systematically surveyed 59 freely available bioinformatics tools for exploring target mutation-induced drug resistance. We analyzed and summarized these resources based on their functionality, data volume, data source, operating principle, performance, etc. And we concisely discussed the strengths, limitations and application examples of these tools. Specifically, we tested some predictive tools and offered some thoughts from the clinician's perspective. Hopefully, this work will provide a useful toolbox for researchers working in the biomedical, pesticide, bioinformatics and pharmaceutical engineering fields, and a good platform for non-specialists to quickly understand drug resistance prediction.

Performances of bioinformatics tools for the analysis of sequencing data of Mycobacterium tuberculosis complex strains

Whole-genome sequencing of Mycobacterium tuberculosis complex (MTBC) strains is a rapidly growing tool to obtain results regarding the resistance and phylogeny of the strains. We evaluated the performances of two bioinformatics tools for the analysis of whole-genome sequences of MTBC strains. Two hundred and twenty-seven MTBC strains were isolated and whole-genome sequenced at the laboratory of Avicenne Hospital between 2015 and 2021. We investigated the resistance and susceptibility status of strains using two online tools, Mykrobe and PhyResSE. We compared the genotypic and phenotypic resistance results obtained by drug susceptibility testing. Unlike with the Mykrobe tool, sequencing quality data were obtained using PhyResSE: average coverage of 98% and average depth of 119X. We found a similar concordance between phenotypic and genotypic results when determining susceptibility to first-line anti-tuberculosis drugs (95%) with both tools. The sensitivity and specificity of each tool compared to the phenotypic method were respectively 72% [52-87] and 98% [96-99] for Mykrobe and 76% [57-90] and 97% [94-99] for PhyResSE. Mykrobe and PhyResSE were easy to use and efficient. These platforms are accessible to people not trained in bioinformatics and constitute a complementary approach to phenotypic methods for the study of MTBC strains.

Consequences of rpoB mutations missed by the GenoType MTBDRplus assay in a programmatic setting in South Africa

Background

Rifampicin resistance missed by commercial rapid molecular assays but detected by phenotypic assays may lead to discordant susceptibility results and affect patient management.

Objective

This study was conducted to evaluate the causes of rifampicin resistance missed by the GenoType MTBDRplus and its impact on the programmatic management of tuberculosis in KwaZulu-Natal, South Africa.

Methods

We analysed routine tuberculosis programme data from January 2014 to December 2014 on isolates showing rifampicin susceptibility on the GenoType MTBDRplus assay but resistance on the phenotypic agar proportion method. Whole-genome sequencing was performed on a subset of these isolates.

Results

Out of 505 patients with isoniazid mono-resistant tuberculosis on the MTBDRplus, 145 (28.7%) isolates showed both isoniazid and rifampicin resistance on the phenotypic assay. The mean time from MTBDRplus results to initiation of drug-resistant tuberculosis therapy was 93.7 days. 65.7% of the patients had received previous tuberculosis treatment. The most common mutations detected in the 36 sequenced isolates were I491F (16; 44.4%) and L452P (12; 33.3%). Among the 36 isolates, resistance to other anti-tuberculosis drugs was 69.4% for pyrazinamide, 83.3% for ethambutol, 69.4% for streptomycin, and 50% for ethionamide.

Conclusion

Missed rifampicin resistance was mostly due to the I491F mutation located outside the MTBDRplus detection area and the L452P mutation, which was not included in the initial version 2 of the MTBDRplus. This led to substantial delays in the initiation of appropriate therapy. The previous tuberculosis treatment history and the high level of resistance to other anti-tuberculosis drugs suggest an accumulation of resistance.

Investigating resistance in clinical Mycobacterium tuberculosis complex isolates with genomic and phenotypic antimicrobial susceptibility testing: a multicentre observational study

BACKGROUND

Whole-genome sequencing (WGS) of Mycobacterium tuberculosis complex has become an important tool in diagnosis and management of drug-resistant tuberculosis. However, data correlating resistance genotype with quantitative phenotypic antimicrobial susceptibility testing (AST) are scarce.

METHODS

In a prospective multicentre observational study, 900 clinical M tuberculosis complex isolates were collected from adults with drug-resistant tuberculosis in five high-endemic tuberculosis settings around the world (Georgia, Moldova, Peru, South Africa, and Viet Nam) between Dec 5, 2014, and Dec 12, 2017. Minimum inhibitory concentrations (MICs) and resulting binary phenotypic AST results for up to nine antituberculosis drugs were determined and correlated with resistance-conferring mutations identified by WGS.

FINDINGS

Considering WHO-endorsed critical concentrations as reference, WGS had high accuracy for prediction of resistance to isoniazid (sensitivity 98·8% [95% CI 98·5-99·0]; specificity 96·6% [95% CI 95·2-97·9]), levofloxacin (sensitivity 94·8% [93·3-97·6]; specificity 97·1% [96·7-97·6]), kanamycin (sensitivity 96·1% [95·4-96·8]; specificity 95·0% [94·4-95·7]), amikacin (sensitivity 97·2% [96·4-98·1]; specificity 98·6% [98·3-98·9]), and capreomycin (sensitivity 93·1% [90·0-96·3]; specificity 98·3% [98·0-98·7]). For rifampicin, pyrazinamide, and ethambutol, the specificity of resistance prediction was suboptimal (64·0% [61·0-67·1], 83·8% [81·0-86·5], and 40·1% [37·4-42·9], respectively). Specificity for rifampicin increased to 83·9% when borderline mutations with MICs overlapping with the critical concentration were excluded. Consequently, we highlighted mutations in M tuberculosis complex isolates that are often falsely identified as susceptible by phenotypic AST, and we identified potential novel resistance-conferring mutations.

INTERPRETATION

The combined analysis of mutations and quantitative phenotypes shows the potential of WGS to produce a refined interpretation of resistance, which is needed for individualised therapy, and eventually could allow differential drug dosing. However, variability of MIC data for some M tuberculosis complex isolates carrying identical mutations also reveals limitations of our understanding of the genotype and phenotype relationships (eg, including epistasis and strain genetic background).

FUNDING

Bill & Melinda Gates Foundation, German Centre for Infection Research, German Research Foundation, Excellence Cluster Precision Medicine of Inflammation (EXC 2167), and Leibniz ScienceCampus EvoLUNG.

Whole genome sequencing analysis to evaluate the influence of T2DM on polymorphisms associated with drug resistance in M. tuberculosis

Background

Type 2 diabetes mellitus (T2DM) has been associated with treatment failure, and the development of drug resistance in tuberculosis (TB). Also, whole-genome sequencing has provided a better understanding and allowed the growth of knowledge about polymorphisms in genes associated with drug resistance. Considering the above, this study analyzes genome sequences to evaluate the influence of type 2 diabetes mellitus in the development of mutations related to tuberculosis drug resistance. M. tuberculosis isolates from individuals with (n = 74), and without (n = 74) type 2 diabetes mellitus was recovered from online repositories, and further analyzed.

Results

The results showed the presence of 431 SNPs with similar proportions between diabetics, and non-diabetics individuals (48% vs. 52%), but with no significant relationship. A greater number of mutations associated with rifampicin resistance was observed in the T2DM-TB individuals (23.2% vs. 16%), and the exclusive presence of rpoBQ432L, rpoBQ432P, rpoBS441L, and rpoBH445L variants. While these variants are not private to T2DM-TB cases they are globally rare highlighting a potential role of T2DM. The phylogenetic analysis showed 12 sublineages, being 4.1.1.3, and 4.1.2.1 the most prevalent in T2DM-TB individuals but not differing from those most prevalent in their geographic location. Four clonal complexes were found, however, no significant relationship with T2DM was observed. Samples size and potential sampling biases prevented us to look for significant associations.

Conclusions

The occurrence of globally rare rifampicin variants identified only in isolates from individuals with T2DM could be due to the hyperglycemic environment within the host. Therefore, further studies about the dynamics of SNPs’ generation associated with antibiotic resistance in patients with diabetes mellitus are necessary.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12864-022-08709-z.

Application of Next Generation Sequencing for Diagnosis and Clinical Management of Drug-Resistant Tuberculosis: Updates on Recent Developments in the Field

The World Health Organization’s End TB Strategy prioritizes universal access to an early diagnosis and comprehensive drug susceptibility testing (DST) for all individuals with tuberculosis (TB) as a key component of integrated, patient-centered TB care. Next generation whole genome sequencing (WGS) and its associated technology has demonstrated exceptional potential for reliable and comprehensive resistance prediction for Mycobacterium tuberculosis isolates, allowing for accurate clinical decisions. This review presents a descriptive analysis of research describing the potential of WGS to accelerate delivery of individualized care, recent advances in sputum-based WGS technology and the role of targeted sequencing for resistance detection. We provide an update on recent research describing the mechanisms of resistance to new and repurposed drugs and the dynamics of mixed infections and its potential implication on TB diagnosis and treatment. Whilst the studies reviewed here have greatly improved our understanding of recent advances in this arena, it highlights significant challenges that remain. The wide-spread introduction of new drugs in the absence of standardized DST has led to rapid emergence of drug resistance. This review highlights apparent gaps in our knowledge of the mechanisms contributing to resistance for these new drugs and challenges that limit the clinical utility of next generation sequencing techniques. It is recommended that a combination of genotypic and phenotypic techniques is warranted to monitor treatment response, curb emerging resistance and further dissemination of drug resistance.

Antimicrobial Resistance in Rivers: A Review of the Genes Detected and New Challenges

River ecosystems are very important parts of the water cycle and an excellent habitat, food, and drinking water source for many organisms, including humans. Antibiotics are emerging contaminants which can enter rivers from various sources. Several antibiotics and their related antibiotic resistance genes (ARGs) have been detected in these ecosystems by various research programs and could constitute a substantial problem. The presence of antibiotics and other resistance cofactors can boost the development of ARGs in the chromosomes or mobile genetic elements of natural bacteria in rivers. The ARGs in environmental bacteria can also be transferred to clinically important pathogens. However, antibiotics and their resistance genes are both not currently monitored by national or international authorities responsible for controlling the quality of water bodies. For example, they are not included in the contaminant list in the European Water Framework Directive or in the US list of Water-Quality Benchmarks for Contaminants. Although ARGs are naturally present in the environment, very few studies have focused on non-impacted rivers to assess the background ARG levels in rivers, which could provide some useful indications for future environmental regulation and legislation. The present study reviews the antibiotics and associated ARGs most commonly measured and detected in rivers, including the primary analysis tools used for their assessment. In addition, other factors that could enhance antibiotic resistance, such as the effects of chemical mixtures, the effects of climate change, and the potential effects of the coronavirus disease 2019 pandemic, are discussed. Environ Toxicol Chem 2022;41:687-714. © 2022 SETAC.

Whole genome sequencing of Mycobacterium tuberculosis for prediction of drug resistance

Whole-genome sequencing (WGS) has shown tremendous potential in rapid diagnosis of drug-resistant tuberculosis (TB). In the current study, we performed WGS on drug-resistant Mycobacterium tuberculosis isolates obtained from Shanghai (n = 137) and Russia (n = 78). We aimed to characterise the underlying and high-frequency novel drug-resistance-conferring mutations, and also create valuable combinations of resistance mutations with high predictive sensitivity to predict multidrug- and extensively drug-resistant tuberculosis (MDR/XDR-TB) phenotype using a bootstrap method. Most strains belonged to L2.2, L4.2, L4.4, L4.5 and L4.8 lineages. We found that WGS could predict 82.07% of phenotypically drug-resistant domestic strains. The prediction sensitivity for rifampicin (RIF), isoniazid (INH), ethambutol (EMB), streptomycin (STR), ofloxacin (OFL), amikacin (AMK) and capreomycin (CAP) was 79.71%, 86.30%, 76.47%, 88.37%, 83.33%, 70.00% and 70.00%, respectively. The mutation combination with the highest sensitivity for MDR prediction was rpoB S450L + rpoB H445A/P + katG S315T + inhA I21T + inhA S94A, with a sensitivity of 92.17% (0.8615, 0.9646), and the mutation combination with highest sensitivity for XDR prediction was rpoB S450L + katG S315T + gyrA D94G + rrs A1401G, with a sensitivity of 92.86% (0.8158, 0.9796). The molecular information presented here will be of particular value for the rapid clinical detection of MDR- and XDR-TB isolates through laboratory diagnosis.

Comprehensive and accurate genetic variant identification from contaminated and low-coverage Mycobacterium tuberculosis whole genome sequencing data

Improved understanding of the genomic variants that allow Mycobacterium tuberculosis (Mtb ) to acquire drug resistance, or tolerance, and increase its virulence are important factors in controlling the current tuberculosis epidemic. Current approaches to Mtb sequencing, however, cannot reveal Mtb ’s full genomic diversity due to the strict requirements of low contamination levels, high Mtb sequence coverage and elimination of complex regions. We have developed the XBS (compleX Bacterial Samples) bioinformatics pipeline, which implements joint calling and machine-learning-based variant filtering tools to specifically improve variant detection in the important Mtb samples that do not meet these criteria, such as those from unbiased sputum samples. Using novel simulated datasets, which permit exact accuracy verification, XBS was compared to the UVP and MTBseq pipelines. Accuracy statistics showed that all three pipelines performed equally well for sequence data that resemble those obtained from culture isolates of high depth of coverage and low-level contamination. In the complex genomic regions, however, XBS accurately identified 9.0 % more SNPs and 8.1 % more single nucleotide insertions and deletions than the WHO-endorsed unified analysis variant pipeline. XBS also had superior accuracy for sequence data that resemble those obtained directly from sputum samples, where depth of coverage is typically very low and contamination levels are high. XBS was the only pipeline not affected by low depth of coverage (5–10×), type of contamination and excessive contamination levels (>50 %). Simulation results were confirmed using whole genome sequencing (WGS) data from clinical samples, confirming the superior performance of XBS with a higher sensitivity (98.8%) when analysing culture isolates and identification of 13.9 % more variable sites in WGS data from sputum samples as compared to MTBseq, without evidence for false positive variants when rRNA regions were excluded. The XBS pipeline facilitates sequencing of less-than-perfect Mtb samples. These advances will benefit future clinical applications of Mtb sequencing, especially WGS directly from clinical specimens, thereby avoiding in vitro biases and making many more samples available for drug resistance and other genomic analyses. The additional genetic resolution and increased sample success rate will improve genome-wide association studies and sequence-based transmission studies.

Evaluation of whole-genome sequence data analysis approaches for short- and long-read sequencing of Mycobacterium tuberculosis

Whole-genome sequencing (WGS) of Mycobacterium tuberculosis (MTB) isolates can be used to get an accurate diagnosis, to guide clinical decision making, to control tuberculosis (TB) and for outbreak investigations. We evaluated the performance of long-read (LR) and/or short-read (SR) sequencing for anti-TB drug-resistance prediction using the TBProfiler and Mykrobe tools, the fraction of genome recovery, assembly accuracies and the robustness of two typing approaches based on core-genome SNP (cgSNP) typing and core-genome multi-locus sequence typing (cgMLST). Most of the discrepancies between phenotypic drug-susceptibility testing (DST) and drug-resistance prediction were observed for the first-line drugs rifampicin, isoniazid, pyrazinamide and ethambutol, mainly with LR sequence data. Resistance prediction to second-line drugs made by both TBProfiler and Mykrobe tools with SR- and LR-sequence data were in complete agreement with phenotypic DST except for one isolate. The SR assemblies were more accurate than the LR assemblies, having significantly (P <0.05) fewer indels and mismatches per 100 kbp. However, the hybrid and LR assemblies had slightly higher genome fractions. For LR assemblies, Canu followed by Racon, and Medaka polishing was the most accurate approach. The cgSNP approach, based on either reads or assemblies, was more robust than the cgMLST approach, especially for LR sequence data. In conclusion, anti-TB drug-resistance prediction, particularly with only LR sequence data, remains challenging, especially for first-line drugs. In addition, SR assemblies appear more accurate than LR ones, and reproducible phylogeny can be achieved using cgSNP approaches.

A bioinformatics pipeline for Mycobacterium tuberculosis sequencing that cleans contaminant reads from sputum samples

Next-Generation Sequencing (NGS) is widely used to investigate genomic variation. In several studies, the genetic variation of Mycobacterium tuberculosis has been analyzed in sputum samples without previous culture, using target enrichment methodologies for NGS. Alignments obtained by different programs generally map the sequences under default parameters, and from these results, it is assumed that only Mycobacterium reads will be obtained. However, variants of interest microorganism in clinical samples can be confused with a vast collection of reads from other bacteria, viruses, and human DNA. Currently, there are no standardized pipelines, and the cleaning success is never verified since there is a lack of rigorous controls to identify and remove reads from other sputum-microorganisms genetically similar to M. tuberculosis. Therefore, we designed a bioinformatic pipeline to process NGS data from sputum samples, including several filters and quality control points to identify and eliminate non-M. tuberculosis reads to obtain a reliable genetic variant report. Our proposal uses the SURPI software as a taxonomic classifier to filter input sequences and perform a mapping that provides the highest percentage of Mycobacterium reads, minimizing the reads from other microorganisms. We then use the filtered sequences to perform variant calling with the GATK software, ensuring the mapping quality, realignment, recalibration, hard-filtering, and post-filter to increase the reliability of the reported variants. Using default mapping parameters, we identified reads of contaminant bacteria, such as Streptococcus, Rhotia, Actinomyces, and Veillonella. Our final mapping strategy allowed a sequence identity of 97.8% between the input reads and the whole M. tuberculosis reference genome H37Rv using a genomic edit distance of three, thus removing 98.8% of the off-target sequences with a Mycobacterium reads loss of 1.7%. Finally, more than 200 unreliable genetic variants were removed during the variant calling, increasing the report’s reliability.

Simultaneous Determination of Mycobacterium leprae Drug Resistance and Single-Nucleotide Polymorphism Genotype by Use of Nested Multiplex PCR with Amplicon Sequencing

Mycobacterium leprae is the predominant cause of leprosy worldwide, and its genotypes can be classified into four single-nucleotide polymorphism (SNP) types and 16 subtypes. Determining M. leprae drug resistance and genotype is typically done by PCR and Sanger DNA sequencing, which require substantial effort. Here, we describe a rapid method involving multiplex PCR in combination with nested amplification and next-generation sequence analysis that allows simultaneous determination of M. leprae drug resistance and SNP genotype directly from clinical specimens. We used this method to analyze clinical samples from two paucibacillary, nine multibacillary, and six type-undetermined leprosy patients. Regions in folP1, rpoB, gyrA, and gyrB that determine drug resistance and those for 84 SNP-InDels in the M. leprae genome were amplified from clinical samples and their sequences determined. The results showed that seven samples were subtype 1A, three were 1D, and seven were 3K. Three samples of the subtype 3K had folp1 mutation. The method may allow more rapid genetic analyses of M. leprae in clinical samples.

INGOT-DR: an interpretable classifier for predicting drug resistance in M. tuberculosis

Motivation

Prediction of drug resistance and identification of its mechanisms in bacteria such as Mycobacterium tuberculosis, the etiological agent of tuberculosis, is a challenging problem. Solving this problem requires a transparent, accurate, and flexible predictive model. The methods currently used for this purpose rarely satisfy all of these criteria. On the one hand, approaches based on testing strains against a catalogue of previously identified mutations often yield poor predictive performance; on the other hand, machine learning techniques typically have higher predictive accuracy, but often lack interpretability and may learn patterns that produce accurate predictions for the wrong reasons. Current interpretable methods may either exhibit a lower accuracy or lack the flexibility needed to generalize them to previously unseen data.

Contribution

In this paper we propose a novel technique, inspired by group testing and Boolean compressed sensing, which yields highly accurate predictions, interpretable results, and is flexible enough to be optimized for various evaluation metrics at the same time.

Results

We test the predictive accuracy of our approach on five first-line and seven second-line antibiotics used for treating tuberculosis. We find that it has a higher or comparable accuracy to that of commonly used machine learning models, and is able to identify variants in genes with previously reported association to drug resistance. Our method is intrinsically interpretable, and can be customized for different evaluation metrics. Our implementation is available at github.com/hoomanzabeti/INGOT_DR and can be installed via The Python Package Index (Pypi) under ingotdr. This package is also compatible with most of the tools in the Scikit-learn machine learning library.

Genomic-based surveillance reveals high ongoing transmission of multi-drug-resistant Mycobacterium tuberculosis in Southern Brazil

Genomic-based surveillance on the occurrence of drug resistance and its transmission dynamics has emerged as a powerful tool for the control of tuberculosis (TB). A whole-genome sequencing approach, phenotypic testing and clinical-epidemiological investigation were used to undertake a retrospective population-based study on drug-resistant (DR)-TB in Rio Grande do Sul, the largest state in Southern Brazil. The analysis included 305 resistant Mycobacterium tuberculosis strains sampled statewide from 2011 to 2014, and covered 75.7% of all DR-TB cases identified in this period. Lineage 4 was found to be predominant (99.3%), with high sublineage-level diversity composed mainly of 4.3.4.2 [Latin American and Mediterranean (LAM)/RD174], 4.3.3 (LAM/RD115) and 4.1.2.1 (Haarlem/RD182) sublineages. Genomic diversity was also reflected in resistance of the variants to first-line drugs. A large number of distinct resistance-conferring mutations, including variants that have not been reported previously in any other setting worldwide, and 22 isoniazid-monoresistant strains with mutations described as disputed in the rpoB gene but causing rifampicin resistance generally missed by automated phenotypic tests as BACTEC MGIT. Using a cut-off of five single nucleotide polymorphisms, the estimated recent transmission rate was 55.1%, with 168 strains grouped into 28 genomic clusters. The most worrying fact concerns multi-drug-resistant (MDR) strains, of which 73.4% were clustered. Different resistance profiles and acquisition of novel mutations intraclusters revealed important amplification of resistance in the region. This study described the diversity of M. tuberculosis strains, the basis of drug resistance, and ongoing transmission dynamics across the largest state in Southern Brazil, stressing the urgent need for MDR-TB transmission control state-wide.

Perspectives for systems biology in the management of tuberculosis

Standardised management of tuberculosis may soon be replaced by individualised, precision medicine-guided therapies informed with knowledge provided by the field of systems biology. Systems biology is a rapidly expanding field of computational and mathematical analysis and modelling of complex biological systems that can provide insights into mechanisms underlying tuberculosis, identify novel biomarkers, and help to optimise prevention, diagnosis and treatment of disease. These advances are critically important in the context of the evolving epidemic of drug-resistant tuberculosis. Here, we review the available evidence on the role of systems biology approaches - human and mycobacterial genomics and transcriptomics, proteomics, lipidomics/metabolomics, immunophenotyping, systems pharmacology and gut microbiomes - in the management of tuberculosis including prediction of risk for disease progression, severity of mycobacterial virulence and drug resistance, adverse events, comorbidities, response to therapy and treatment outcomes. Application of the Grading of Recommendations, Assessment, Development and Evaluation (GRADE) approach demonstrated that at present most of the studies provide "very low" certainty of evidence for answering clinically relevant questions. Further studies in large prospective cohorts of patients, including randomised clinical trials, are necessary to assess the applicability of the findings in tuberculosis prevention and more efficient clinical management of patients.

A Bioinformatics Whole-Genome Sequencing Workflow for Clinical Mycobacterium tuberculosis Complex Isolate Analysis, Validated Using a Reference Collection Extensively Characterized with Conventional Methods and In Silico Approaches

The use of whole-genome sequencing (WGS) for routine typing of bacterial isolates has increased substantially in recent years. For Mycobacterium tuberculosis (MTB), in particular, WGS has the benefit of drastically reducing the time required to generate results compared to most conventional phenotypic methods. Consequently, a multitude of solutions for analyzing WGS MTB data have been developed, but their successful integration in clinical and national reference laboratories is hindered by the requirement for their validation, for which a consensus framework is still largely absent. We developed a bioinformatics workflow for (Illumina) WGS-based routine typing of MTB complex (MTBC) member isolates allowing complete characterization, including (sub)species confirmation and identification (16S, csb/RD, hsp65), single nucleotide polymorphism (SNP)-based antimicrobial resistance (AMR) prediction, and pathogen typing (spoligotyping, SNP barcoding, and core genome multilocus sequence typing). Workflow performance was validated on a per-assay basis using a collection of 238 in-house-sequenced MTBC isolates, extensively characterized with conventional molecular biology-based approaches supplemented with public data. For SNP-based AMR prediction, results from molecular genotyping methods were supplemented with in silico modified data sets, allowing us to greatly increase the set of evaluated mutations. The workflow demonstrated very high performance with performance metrics of >99% for all assays, except for spoligotyping, where sensitivity dropped to ∼90%. The validation framework for our WGS-based bioinformatics workflow can aid in the standardization of bioinformatics tools by the MTB community and other SNP-based applications regardless of the targeted pathogen(s). The bioinformatics workflow is available for academic and nonprofit use through the Galaxy instance of our institute at https://galaxy.sciensano.be.

Evaluation of drug susceptibility profile of Mycobacterium tuberculosis Lineage 1 from Brazil based on whole genome sequencing and phenotypic methods

BACKGROUND

The evaluation of procedures for drug susceptibility prediction of Mycobacterium tuberculosis based on genomic data against the conventional reference method test based on culture is realistic considering the scenario of growing number of tools proposals based on whole-genome sequences (WGS).

OBJECTIVES

This study aimed to evaluate drug susceptibility testing (DST) outcome based on WGS tools and the phenotypic methods performed on isolates of M. tuberculosis Lineage 1 from the state of Pará, Brazil, generally associated with low levels of drug resistance.

METHODOLOGY

Culture based DST was performed using the Proportion Method in Löwenstein-Jensen medium on 71 isolates that had been submitted to WGS. We analysed the seven main genome sequence-based tools for resistance and lineage prediction applied to M. tuberculosis and for comparison evaluation we have used the Kappa concordance test.

FINDINGS

When comparing the WGS-based tools against the DST, we observed the highest level of agreement using TB-profiler. Among the tools, TB-profiler, KvarQ and Mykrobe were those which identified the largest number of TB-MDR cases. Comparing the four most sensitive tools regarding resistance prediction, agreement was observed for 43 genomes.

MAIN CONCLUSIONS

Drug resistance profiling using next-generation sequencing offers rapid assessment of resistance-associated mutations, therefore facilitating rapid access to effective treatment.

Assessing Nanopore Sequencing for Clinical Diagnostics: a Comparison of Next-Generation Sequencing (NGS) Methods for Mycobacterium tuberculosis

Next-generation sequencing technologies are being rapidly adopted as a tool of choice for diagnostic and outbreak investigation in public health laboratories. However, costs of operation and the need for specialized staff remain major hurdles for laboratories with limited resources for implementing these technologies. This project aimed to assess the feasibility of using Oxford Nanopore MinION whole-genome sequencing data of Mycobacterium tuberculosis isolates for species identification, in silico spoligotyping, detection of mutations associated with antimicrobial resistance (AMR) to accurately predict drug susceptibility profiles, and phylogenetic analysis to detect transmission between cases. The results were compared prospectively in real time to those obtained with our current clinically validated Illumina MiSeq sequencing assay for M. tuberculosis and phenotypic drug susceptibility testing results when available. Our assessment of 431 sequenced samples over a 32-week period demonstrates that, when using the proper quality controls and thresholds, the MinION can achieve levels of genotyping analysis and phenotypic resistance predictions comparable to those of the Illumina MiSeq at a very competitive cost per sample. Our results indicate that nanopore sequencing can be a suitable alternative to, or complement, currently used sequencing platforms in a clinical setting and has the potential to be widely adopted in public health laboratories in the near future.

Prediction of rifampicin resistance beyond the RRDR using structure-based machine learning approaches

Rifampicin resistance is a major therapeutic challenge, particularly in tuberculosis, leprosy, P. aeruginosa and S. aureus infections, where it develops via missense mutations in gene rpoB. Previously we have highlighted that these mutations reduce protein affinities within the RNA polymerase complex, subsequently reducing nucleic acid affinity. Here, we have used these insights to develop a computational rifampicin resistance predictor capable of identifying resistant mutations even outside the well-defined rifampicin resistance determining region (RRDR), using clinical M. tuberculosis sequencing information. Our tool successfully identified up to 90.9% of M. tuberculosis rpoB variants correctly, with sensitivity of 92.2%, specificity of 83.6% and MCC of 0.69, outperforming the current gold-standard GeneXpert-MTB/RIF. We show our model can be translated to other clinically relevant organisms: M. leprae, P. aeruginosa and S. aureus, despite weak sequence identity. Our method was implemented as an interactive tool, SUSPECT-RIF (StrUctural Susceptibility PrEdiCTion for RIFampicin), freely available at https://biosig.unimelb.edu.au/suspect_rif/ .

Perspective for Precision Medicine for Tuberculosis

Tuberculosis is a bacterial infectious disease that is mainly transmitted from human to human via infectious aerosols. Currently, tuberculosis is the leading cause of death by an infectious disease world-wide. In the past decade, the number of patients affected by tuberculosis has increased by ~20 percent and the emergence of drug-resistant strains of Mycobacterium tuberculosis challenges the goal of elimination of tuberculosis in the near future. For the last 50 years, management of patients with tuberculosis has followed a standardized management approach. This standardization neglects the variation in human susceptibility to infection, immune response, the pharmacokinetics of drugs, and the individual duration of treatment needed to achieve relapse-free cure. Here we propose a package of precision medicine-guided therapies that has the prospect to drive clinical management decisions, based on both host immunity and M. tuberculosis strains genetics. Recently, important scientific discoveries and technological advances have been achieved that provide a perspective for individualized rather than standardized management of patients with tuberculosis. For the individual selection of best medicines and host-directed therapies, personalized drug dosing, and treatment durations, physicians treating patients with tuberculosis will be able to rely on these advances in systems biology and to apply them at the bedside.

Transmission of Mycobacterium tuberculosis presenting unusually high discordance between genotypic and phenotypic resistance to rifampicin in an endemic tuberculosis setting

BACKGROUND

Since the implementation of the Xpert MTB/RIF in Sao Paulo, Brazil, numerous Mycobacterium tuberculosis isolates presenting "rifampicin-resistant genotype with rifampicin-susceptible phenotype" were observed.

OBJECTIVE

To evaluate the prevalence, rpoB mutations and transmission of M. tuberculosis resistant to rifampicin on Xpert MTB/RIF but susceptible on BACTEC MGIT system, in Sao Paulo state.

METHODS

Patients' isolates with this pattern of rifampicin discordance, collected from 2014 to 2017, had their rpoB predominant rifampicin-resistance-determining region sequenced and were genotyped by IS6110 restriction fragment-length polymorphism.

FINDINGS

The prevalence of rifampicin-discordant M. tuberculosis with genotypic resistance was 55.1% (156/283). Among the sequenced and genotyped isolates, 75.5% (111/147) were in clusters, largely associated with the type of rpoB mutation. Most isolates (98.6%; 72/73) harbouring the predominant mutation, His445Asn, were pooled into the two largest clusters, SP2ga (42/72; 58.3%) and SP5o (12/72; 16.7%). Ranking second, isolates carrying the silent mutation Phe433Phe were mostly (92.3%; 24/26) gathered into four groups of the family SP25.

CONCLUSION

These findings suggest that this unusual high rifampicin discrepancy proportion was greatly influenced by few actively circulating clusters. Further studies on many of the rpoB mutations identified in our setting are needed to elucidate their association with phenotypic rifampicin resistance.

Prediction of Mycobacterium tuberculosis pyrazinamidase function based on structural stability, physicochemical and geometrical descriptors

BACKGROUND

Pyrazinamide is an important drug against the latent stage of tuberculosis and is used in both first- and second-line treatment regimens. Pyrazinamide-susceptibility test usually takes a week to have a diagnosis to guide initial therapy, implying a delay in receiving appropriate therapy. The continued increase in multi-drug resistant tuberculosis and the prevalence of pyrazinamide resistance in several countries makes the development of assays for prompt identification of resistance necessary. The main cause of pyrazinamide resistance is the impairment of pyrazinamidase function attributed to mutations in the promoter and/or pncA coding gene. However, not all pncA mutations necessarily affect the pyrazinamidase function.

OBJECTIVE

To develop a methodology to predict pyrazinamidase function from detected mutations in the pncA gene.

METHODS

We measured the catalytic constant (kcat), KM, enzymatic efficiency, and enzymatic activity of 35 recombinant mutated pyrazinamidase and the wild type (Protein Data Bank ID = 3pl1). From all the 3D modeled structures, we extracted several predictors based on three categories: structural stability (estimated by normal mode analysis and molecular dynamics), physicochemical, and geometrical characteristics. We used a stepwise Akaike's information criterion forward multiple log-linear regression to model each kinetic parameter with each category of predictors. We also developed weighted models combining the three categories of predictive models for each kinetic parameter. We tested the robustness of the predictive ability of each model by 6-fold cross-validation against random models.

RESULTS

The stability, physicochemical, and geometrical descriptors explained most of the variability (R2) of the kinetic parameters. Our models are best suited to predict kcat, efficiency, and activity based on the root-mean-square error of prediction of the 6-fold cross-validation.

CONCLUSIONS

This study shows a quick approach to predict the pyrazinamidase function only from the pncA sequence when point mutations are present. This can be an important tool to detect pyrazinamide resistance.

Whole-genome sequencing and Mycobacterium tuberculosis: Challenges in sample preparation and sequencing data analysis

The numbers of patients with tuberculosis (TB) caused by resistant strains are still alarming. Therefore, it is necessary to determine resistance more quickly and precisely, than it is with the currently used phenotypic and genotypic methods. In recent years, technological advances have been made and the whole-genome sequencing (WGS) method has been introduced as a part of routine diagnostics in clinical laboratories. Comparing a wide range of mycobacterial genomic variations with a reference genome leads to a consistent evaluation of molecular-epidemiology and resistance of Mycobacterium tuberculosis (M. tuberculosis) to a wide range of anti-TB drugs. The quality of the obtained sequencing data is closely related to the type of sample and the method used for DNA extraction and sequencing library preparation. Moreover, the correct interpretation of results is also influenced by a bioinformatic data processing. A large number of bioinformatics pipelines are currently available, the sensitivity of which varies due to the different sizes of databases containing relevant mutations. This review focuses on the individual steps included in the sequencing workflow and factors that may affect the interpretation of final results.

Genotyping, Assessment of Virulence and Antibacterial Resistance of the Rostov Strain of Mycobacterium tuberculosis Attributed to the Central Asia Outbreak Clade

The Central Asia Outbreak (CAO) clade is a growing public health problem for Central Asian countries. Members of the clade belong to the narrow branch of the Mycobacterium tuberculosis Beijing genotype and are characterized by multidrug resistance and increased transmissibility. The Rostov strain of M. tuberculosis isolated in Russia and attributed to the CAO clade based on PCR-assay and whole genome sequencing and the laboratory strain H37Rv were selected to evaluate the virulence on C57Bl/6 mice models by intravenous injection. All mice infected with the Rostov strain succumbed to death within a 48-day period, while more than half of the mice infected by the H37Rv strain survived within a 90-day period. Mice weight analysis revealed irreversible and severe depletion of animals infected with the Rostov strain compared to H37Rv. The histological investigation of lung and liver tissues of mice on the 30th day after injection of mycobacterial bacilli showed that the pattern of pathological changes generated by two strains were different. Moreover, bacterial load in the liver and lungs was higher for the Rostov strain infection. In conclusion, our data demonstrate that the drug-resistant Rostov strain exhibits a highly virulent phenotype which can be partly explained by the CAO-specific mutations.

Multi-Label Random Forest Model for Tuberculosis Drug Resistance Classification and Mutation Ranking

Resistance prediction and mutation ranking are important tasks in the analysis of Tuberculosis sequence data. Due to standard regimens for the use of first-line antibiotics, resistance co-occurrence, in which samples are resistant to multiple drugs, is common. Analysing all drugs simultaneously should therefore enable patterns reflecting resistance co-occurrence to be exploited for resistance prediction. Here, multi-label random forest (MLRF) models are compared with single-label random forest (SLRF) for both predicting phenotypic resistance from whole genome sequences and identifying important mutations for better prediction of four first-line drugs in a dataset of 13402 Mycobacterium tuberculosis isolates. Results confirmed that MLRFs can improve performance compared to conventional clinical methods (by 18.10%) and SLRFs (by 0.91%). In addition, we identified a list of candidate mutations that are important for resistance prediction or that are related to resistance co-occurrence. Moreover, we found that retraining our analysis to a subset of top-ranked mutations was sufficient to achieve satisfactory performance. The source code can be found at http://www.robots.ox.ac.uk/~davidc/code.php.

Systematic Review of Whole-Genome Sequencing Data To Predict Phenotypic Drug Resistance and Susceptibility in Swedish Mycobacterium tuberculosis Isolates, 2016 to 2018

In this retrospective study, whole-genome sequencing (WGS) data generated on an Ion Torrent platform was used to predict phenotypic drug resistance profiles for first- and second-line drugs among Swedish clinical Mycobacterium tuberculosis isolates from 2016 to 2018. The accuracy was ∼99% for all first-line drugs and 100% for four second-line drugs. Our analysis supports the introduction of WGS into routine diagnostics, which might, at least in Sweden, replace phenotypic drug susceptibility testing in the future.

Metabolic Changes of Mycobacterium tuberculosis during the Anti-Tuberculosis Therapy

Tuberculosis, caused by Mycobacterium tuberculosis complex bacteria, remains one of the most pressing health problems. Despite the general trend towards reduction of the disease incidence rate, the situation remains extremely tense due to the distribution of the resistant forms. Most often, these strains emerge through the intra-host microevolution of the pathogen during treatment failure. In the present study, the focus was on three serial clinical isolates of Mycobacterium tuberculosis Beijing B0/W148 cluster from one patient with pulmonary tuberculosis, to evaluate their changes in metabolism during anti-tuberculosis therapy. Using whole genome sequencing (WGS), 9 polymorphisms were determined, which occurred in a stepwise or transient manner during treatment and were linked to the resistance (GyrA D94A; inhA t-8a) or virulence. The effect of the inhA t-8a mutation was confirmed on both proteomic and transcriptomic levels. Additionally, the amount of RpsL protein, which is a target of anti-tuberculosis drugs, was reduced. At the systemic level, profound changes in metabolism, linked to the evolution of the pathogen in the host and the effects of therapy, were documented. An overabundance of the FAS-II system proteins (HtdX, HtdY) and expression changes in the virulence factors have been observed at the RNA and protein levels.

Resistance Sniffer: An online tool for prediction of drug resistance patterns of Mycobacterium tuberculosis isolates using next generation sequencing data

The effective control of multidrug resistant tuberculosis (MDR-TB) relies upon the timely diagnosis and correct treatment of all tuberculosis cases. Whole genome sequencing (WGS) has great potential as a method for the rapid diagnosis of drug resistant Mycobacterium tuberculosis (Mtb) isolates. This method overcomes most of the problems that are associated with current phenotypic drug susceptibility testing. However, the application of WGS in the clinical setting has been deterred by data complexities and skill requirements for implementing the technologies as well as clinical interpretation of the next generation sequencing (NGS) data. The proposed diagnostic application was drawn upon recent discoveries of patterns of Mtb clade-specific genetic polymorphisms associated with antibiotic resistance. A catalogue of genetic determinants of resistance to thirteen anti-TB drugs for each phylogenetic clade was created. A computational algorithm for the identification of states of diagnostic polymorphisms was implemented as an online software tool, Resistance Sniffer (http://resistance-sniffer.bi.up.ac.za/), and as a stand-alone software tool to predict drug resistance in Mtb isolates using complete or partial genome datasets in different file formats including raw Illumina fastq read files. The program was validated on sequenced Mtb isolates with data on antibiotic resistance trials available from GMTV database and from the TB Platform of South African Medical Research Council (SAMRC), Pretoria. The program proved to be suitable for probabilistic prediction of drug resistance profiles of individual strains and large sequence data sets.

System OMICs analysis of Mycobacterium tuberculosis Beijing B0/W148 cluster

Mycobacterium tuberculosis Beijing B0/W148 is one of the most widely distributed clusters in the Russian Federation and in some countries of the former Soviet Union. Recent studies have improved our understanding of the reasons for the "success" of the cluster but this area remains incompletely studied. Here, we focused on the system omics analysis of the RUS_B0 strain belonging to the Beijing B0/W148 cluster. Completed genome sequence of RUS_B0 (CP020093.1) and a collection of WGS for 394 cluster strains were used to describe the main genetic features of the population. In turn, proteome and transcriptome studies allowed to confirm the genomic data and to identify a number of finds that have not previously been described. Our results demonstrated that expression of the whiB6 which contains cluster-specific polymorphism (a151c) increased almost 40 times in RUS_B0. Additionally, the level of ethA transcripts in RUS_B0 was increased by more than 7 times compared to the H37Rv. Start sites for 10 genes were corrected based on the combination of proteomic and transcriptomic data. Additionally, based on the omics approach, we identified 5 new genes. In summary, our analysis allowed us to summarize the available results and also to obtain fundamentally new data.

Antibiotic resistance prediction for Mycobacterium tuberculosis from genome sequence data with Mykrobe

Two billion people are infected with Mycobacterium tuberculosis, leading to 10 million new cases of active tuberculosis and 1.5 million deaths annually. Universal access to drug susceptibility testing (DST) has become a World Health Organization priority. We previously developed a software tool, Mykrobe predictor, which provided offline species identification and drug resistance predictions for M. tuberculosis from whole genome sequencing (WGS) data. Performance was insufficient to support the use of WGS as an alternative to conventional phenotype-based DST, due to mutation catalogue limitations.  Here we present a new tool, Mykrobe, which provides the same functionality based on a new software implementation. Improvements include i) an updated mutation catalogue giving greater sensitivity to detect pyrazinamide resistance, ii) support for user-defined resistance catalogues, iii) improved identification of non-tuberculous mycobacterial species, and iv) an updated statistical model for Oxford Nanopore Technologies sequencing data. Mykrobe is released under MIT license at https://github.com/mykrobe-tools/mykrobe. We incorporate mutation catalogues from the CRyPTIC consortium et al. (2018) and from Walker et al. (2015), and make improvements based on performance on an initial set of 3206 and an independent set of 5845 M. tuberculosis Illumina sequences. To give estimates of error rates, we use a prospectively collected dataset of 4362 M. tuberculosis isolates. Using culture based DST as the reference, we estimate Mykrobe to be 100%, 95%, 82%, 99% sensitive and 99%, 100%, 99%, 99% specific for rifampicin, isoniazid, pyrazinamide and ethambutol resistance prediction respectively. We benchmark against four other tools on 10207 (=5845+4362) samples, and also show that Mykrobe gives concordant results with nanopore data.  We measure the ability of Mykrobe-based DST to guide personalized therapeutic regimen design in the context of complex drug susceptibility profiles, showing 94% concordance of implied regimen with that driven by phenotypic DST, higher than all other benchmarked tools.

Whole genome sequencing for drug-resistant tuberculosis management in South Africa: What gaps would this address and what are the challenges to implementation?

Global control of tuberculosis (TB) has been seriously impacted by the emergence and transmission of its drug-resistant forms. Delayed detection and incomplete characterisation of drug-resistant tuberculosis (DR-TB) contributes to morbidity and mortality, and to ongoing transmission of drug-resistant strains. Current culture-based and molecular diagnostic tools for TB present numerous disadvantages that could potentially lead to misdiagnosis, inappropriate treatment initiation and the amplification of drug resistance. The detection of drug-resistant tuberculosis (DR-TB) in South Africa relies on molecular diagnostic assays such as the Xpert MTB/RIF and line probe assays (MTBDRplus and MTBDRsl). However, these molecular assays are limited to detecting resistance to only a few first-line and second-line drugs. It is for this reason that next-generation sequencing (NGS) and bioinformatics pipelines have been developed for rapid detection of M. tuberculosis drug resistance, with the added advantage that sequence data could also have public health applications through understanding transmission patterns. This review highlights some of the challenges that are currently hampering the diagnosis and control of DR-TB in a high burden setting of the KwaZulu-Natal (KZN) province in South Africa. Shortfalls of current diagnostic techniques for DR-TB are discussed in detail and we also propose how these might be overcome with an accurate and rapid NGS system.

Improved Resistance Prediction in Mycobacterium tuberculosis by Better Handling of Insertions and Deletions, Premature Stop Codons, and Filtering of Non-informative Sites

Resistance in Mycobacterium tuberculosis is a major obstacle for effective treatment of tuberculosis. Multiple studies have shown promising results for predicting drug resistance in M. tuberculosis based on whole genome sequencing (WGS) data, however, these tools are often limited to this single species. We have previously developed a common platform for resistance prediction in multiple species. This platform detects acquired resistance genes (ResFinder) and species-specific chromosomal mutations (PointFinder) associated with resistance, all based on WGS data. In this study, we present a new version of PointFinder together with an updated M. tuberculosis database. PointFinder now includes predictions based on insertions and deletions, and it explicitly reports frameshift mutations and premature stop codons. We found that premature stop codons in four resistance-associated genes (katG, ethA, pncA, and gidB) were over-represented in resistant strains, and we saw an increased prediction performance when including premature stop codons in these genes as resistance markers. Different M. tuberculosis resistance prediction tools vary in performance mostly due to the mutation library used. We found that a well-established mutation library included non-predictive linage markers, and through forward feature selection we eliminated those from the mutation library. Compared to other similar web-based tools, PointFinder performs equally good. The advantages of PointFinder is that together with ResFinder it serves as a common web-based and downloadable platform for resistance detection in multiple species. It is easy to use for clinicians and already widely used in the research community.

Mycobacterium tuberculosis Drug Resistance and Transmission among Human Immunodeficiency Virus-Infected Patients in Ho Chi Minh City, Vietnam

Vietnam has a high burden of tuberculosis (TB) and multidrug-resistant (MDR) TB, but drug resistance patterns and TB transmission dynamics among TB/human immunodeficiency virus (HIV) coinfected patients are not well described. We characterized 200 Mycobacterium tuberculosis isolates from TB/HIV coinfected patients diagnosed at the main TB referral hospital in Ho Chi Minh City, Vietnam. Phenotypic drug susceptibility testing (DST) for first-line drugs, spoligotyping, and 24-locus mycobacterial interspersed repetitive unit (MIRU-24) analysis was performed on all isolates. The 24-locus mycobacterial interspersed repetitive unit clusters and MDR isolates were subjected to whole genome sequencing (WGS). Most of the TB/HIV coinfected patients were young (162/174; 93.1% aged < 45 years) males (173; 86.5% male). Beijing (98; 49.0%) and Indo-Oceanic (70; 35.0%) lineage strains were most common. Phenotypic drug resistance was detected in 84 (42.0%) isolates, of which 17 (8.5%) were MDR; three additional MDR strains were identified on WGS. Strain clustering was reduced from 84.0% with spoligotyping to 20.0% with MIRU-24 typing and to 13.5% with WGS. Whole genome sequencing identified five additional clusters, or members of clusters, not recognized by MIRU-24. In total, 13 small (two to three member) WGS clusters were identified, with less clustering among drug susceptible (2/27; 7.4%) than among drug-resistant strains (25/27; 92.6%). On phylogenetic analysis, strains from TB/HIV coinfected patients were interspersed among strains from the general community; no major clusters indicating transmission among people living with HIV were detected. Tuberculosis/HIV coinfection in Vietnam was associated with high rates of drug resistance and limited genomic evidence of ongoing M. tuberculosis transmission among HIV-infected patients.

Deciphering drug resistance in Mycobacterium tuberculosis using whole-genome sequencing: progress, promise, and challenges

Tuberculosis (TB) is a global infectious threat that is intensified by an increasing incidence of highly drug-resistant disease. Whole-genome sequencing (WGS) studies of Mycobacterium tuberculosis, the causative agent of TB, have greatly increased our understanding of this pathogen. Since the first M. tuberculosis genome was published in 1998, WGS has provided a more complete account of the genomic features that cause resistance in populations of M. tuberculosis, has helped to fill gaps in our knowledge of how both classical and new antitubercular drugs work, and has identified specific mutations that allow M. tuberculosis to escape the effects of these drugs. WGS studies have also revealed how resistance evolves both within an individual patient and within patient populations, including the important roles of de novo acquisition of resistance and clonal spread. These findings have informed decisions about which drug-resistance mutations should be included on extended diagnostic panels. From its origins as a basic science technique, WGS of M. tuberculosis is becoming part of the modern clinical microbiology laboratory, promising rapid and improved detection of drug resistance, and detailed and real-time epidemiology of TB outbreaks. We review the successes and highlight the challenges that remain in applying WGS to improve the control of drug-resistant TB through monitoring its evolution and spread, and to inform more rapid and effective diagnostic and therapeutic strategies.

Application of machine learning techniques to tuberculosis drug resistance analysis

MOTIVATION

Timely identification of Mycobacterium tuberculosis (MTB) resistance to existing drugs is vital to decrease mortality and prevent the amplification of existing antibiotic resistance. Machine learning methods have been widely applied for timely predicting resistance of MTB given a specific drug and identifying resistance markers. However, they have been not validated on a large cohort of MTB samples from multi-centers across the world in terms of resistance prediction and resistance marker identification. Several machine learning classifiers and linear dimension reduction techniques were developed and compared for a cohort of 13 402 isolates collected from 16 countries across 6 continents and tested 11 drugs.

RESULTS

Compared to conventional molecular diagnostic test, area under curve of the best machine learning classifier increased for all drugs especially by 23.11%, 15.22% and 10.14% for pyrazinamide, ciprofloxacin and ofloxacin, respectively (P < 0.01). Logistic regression and gradient tree boosting found to perform better than other techniques. Moreover, logistic regression/gradient tree boosting with a sparse principal component analysis/non-negative matrix factorization step compared with the classifier alone enhanced the best performance in terms of F1-score by 12.54%, 4.61%, 7.45% and 9.58% for amikacin, moxifloxacin, ofloxacin and capreomycin, respectively, as well increasing area under curve for amikacin and capreomycin. Results provided a comprehensive comparison of various techniques and confirmed the application of machine learning for better prediction of the large diverse tuberculosis data. Furthermore, mutation ranking showed the possibility of finding new resistance/susceptible markers.

AVAILABILITY AND IMPLEMENTATION

The source code can be found at http://www.robots.ox.ac.uk/ davidc/code.php.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

Integrating informatics tools and portable sequencing technology for rapid detection of resistance to anti-tuberculous drugs

Background

Mycobacterium tuberculosis resistance to anti-tuberculosis drugs is a major threat to global public health. Whole genome sequencing (WGS) is rapidly gaining traction as a diagnostic tool for clinical tuberculosis settings. To support this informatically, previous work led to the development of the widely used TBProfiler webtool, which predicts resistance to 14 drugs from WGS data. However, for accurate and rapid high throughput of samples in clinical or epidemiological settings, there is a need for a stand-alone tool and the ability to analyse data across multiple WGS platforms, including Oxford Nanopore MinION.

Results

We present a new command line version of the TBProfiler webserver, which includes hetero-resistance calling and will facilitate the batch processing of samples. The TBProfiler database has been expanded to incorporate 178 new markers across 16 anti-tuberculosis drugs. The predictive performance of the mutation library has been assessed using > 17,000 clinical isolates with WGS and laboratory-based drug susceptibility testing (DST) data. An integrated MinION analysis pipeline was assessed by performing WGS on 34 replicates across 3 multi-drug resistant isolates with known resistance mutations. TBProfiler accuracy varied by individual drug. Assuming DST as the gold standard, sensitivities for detecting multi-drug-resistant TB (MDR-TB) and extensively drug-resistant TB (XDR-TB) were 94% (95%CI 93–95%) and 83% (95%CI 79–87%) with specificities of 98% (95%CI 98–99%) and 96% (95%CI 95–97%) respectively. Using MinION data, only one resistance mutation was missed by TBProfiler, involving an insertion in the tlyA gene coding for capreomycin resistance. When compared to alternative platforms (e.g. Mykrobe predictor TB, the CRyPTIC library), TBProfiler demonstrated superior predictive performance across first- and second-line drugs.

Conclusions

The new version of TBProfiler can rapidly and accurately predict anti-TB drug resistance profiles across large numbers of samples with WGS data. The computing architecture allows for the ability to modify the core bioinformatic pipelines and outputs, including the analysis of WGS data sourced from portable technologies. TBProfiler has the potential to be integrated into the point of care and WGS diagnostic environments, including in resource-poor settings.

Electronic supplementary material

The online version of this article (10.1186/s13073-019-0650-x) contains supplementary material, which is available to authorized users.

A large scale evaluation of TBProfiler and Mykrobe for antibiotic resistance prediction in Mycobacterium tuberculosis

Recent years saw a growing interest in predicting antibiotic resistance from whole-genome sequencing data, with promising results obtained for Staphylococcus aureus and Mycobacterium tuberculosis. In this work, we gathered 6,574 sequencing read datasets of M. tuberculosis public genomes with associated antibiotic resistance profiles for both first and second-line antibiotics. We performed a systematic evaluation of TBProfiler and Mykrobe, two widely recognized softwares allowing to predict resistance in M. tuberculosis. The size of the dataset allowed us to obtain confident estimations of their overall predictive performance, to assess precisely the individual predictive power of the markers they rely on, and to study in addition how these softwares behave across the major M. tuberculosis lineages. While this study confirmed the overall good performance of these tools, it revealed that an important fraction of the catalog of mutations they embed is of limited predictive power. It also revealed that these tools offer different sensitivity/specificity trade-offs, which is mainly due to the different sets of mutation they embed but also to their underlying genotyping pipelines. More importantly, it showed that their level of predictive performance varies greatly across lineages for some antibiotics, therefore suggesting that the predictions made by these softwares should be deemed more or less confident depending on the lineage inferred and the predictive performance of the marker(s) actually detected. Finally, we evaluated the relevance of machine learning approaches operating from the set of markers detected by these softwares and show that they present an attractive alternative strategy, allowing to reach better performance for several drugs while significantly reducing the number of candidate mutations to consider.

Comparisons of whole-genome sequencing and phenotypic drug susceptibility testing for Mycobacterium tuberculosis causing MDR-TB and XDR-TB in Thailand

Drug-resistant tuberculosis (TB) is a major public health problem. There is little information regarding the genotypic-phenotypic association of anti-TB drugs, especially for second-line drugs. This study compared phenotypic drug susceptibility testing (DST) with predictions based on whole-genome sequencing (WGS) data for 266 Mycobacterium tuberculosis isolates. Phenotypic DST used the standard proportional method. Clinical isolates of M. tuberculosis collected in Thailand between 1998 and 2013 comprised 51 drug-sensitive strains, six mono-resistant strains, two multiple-resistant strains, 88 multi-drug-resistant strains, 95 pre-extensively drug-resistant strains and 24 extensively drug-resistant strains. WGS analysis was performed using the computer programs PhyResSE and TB-Profiler. TB-Profiler had higher average concordance with phenotypic DST than PhyResSE for both first-line (91.96% vs. 91.4%) and second-line (79.67% vs. 78.20%) anti-TB drugs. The average sensitivity for all anti-TB drugs was also higher (83.13% vs. 72.08%) with slightly lower specificity (83.50% vs. 86.68%). Regardless of the program used, isoniazid, rifampicin and amikacin had the highest concordance with phenotypic DST (96.2%, 93.5% and 95.6%, respectively). Ethambutol, ethionamide and fluoroquinolones had the lowest concordance (87.34%, 81.44% and 73.85%, respectively). Concordance rates of ofloxacin (a second-generation fluoroquinolone), levofloxacin, moxifloxacin and gatifloxacin (third- and fourth-generation fluoroquinolones) were 91.79%, 76.62%, 72.64% and 57.35%, respectively. Discordance between phenotypic and WGS-based DSTs may be due, in part, to the choice of critical concentration and variable reproducibility of the phenotypic tests. It may also be due to limitations of the mutation databases (especially for the second-line drugs) and the analysis program used. Mutations related to fluoroquinolone resistance, especially the later generations, need to be identified.

Deciphering Within-Host Microevolution of Mycobacterium tuberculosis through Whole-Genome Sequencing: the Phenotypic Impact and Way Forward

The Mycobacterium tuberculosis genome is more heterogenous and less genetically stable within the host than previously thought. Currently, only limited data exist on the within-host microevolution, diversity, and genetic stability of M. tuberculosis As a direct consequence, our ability to infer M. tuberculosis transmission chains and to understand the full complexity of drug resistance profiles in individual patients is limited. Furthermore, apart from the acquisition of certain drug resistance-conferring mutations, our knowledge on the function of genetic variants that emerge within a host and their phenotypic impact remains scarce. We performed a systematic literature review of whole-genome sequencing studies of serial and parallel isolates to summarize the knowledge on genetic diversity and within-host microevolution of M. tuberculosis We identified genomic loci of within-host emerged variants found across multiple studies and determined their functional relevance. We discuss important remaining knowledge gaps and finally make suggestions on the way forward.

Predicting Antibiotic Resistance in Gram-Negative Bacilli from Resistance Genes

We developed a rapid high-throughput PCR test and evaluated highly antibiotic-resistant clinical isolates of Escherichia coli (n = 2,919), Klebsiella pneumoniae (n = 1,974), Proteus mirabilis (n = 1,150), and Pseudomonas aeruginosa (n = 1,484) for several antibiotic resistance genes for comparison with phenotypic resistance across penicillins, cephalosporins, carbapenems, aminoglycosides, trimethoprim-sulfamethoxazole, fluoroquinolones, and macrolides. The isolates originated from hospitals in North America (34%), Europe (23%), Asia (13%), South America (12%), Africa (7%), or Oceania (1%) or were of unknown origin (9%).

We developed a rapid high-throughput PCR test and evaluated highly antibiotic-resistant clinical isolates of Escherichia coli (n = 2,919), Klebsiella pneumoniae (n = 1,974), Proteus mirabilis (n = 1,150), and Pseudomonas aeruginosa (n = 1,484) for several antibiotic resistance genes for comparison with phenotypic resistance across penicillins, cephalosporins, carbapenems, aminoglycosides, trimethoprim-sulfamethoxazole, fluoroquinolones, and macrolides. The isolates originated from hospitals in North America (34%), Europe (23%), Asia (13%), South America (12%), Africa (7%), or Oceania (1%) or were of unknown origin (9%). We developed statistical methods to predict phenotypic resistance from resistance genes for 49 antibiotic-organism combinations, including gentamicin, tobramycin, ciprofloxacin, levofloxacin, trimethoprim-sulfamethoxazole, ertapenem, imipenem, cefazolin, cefepime, cefotaxime, ceftazidime, ceftriaxone, ampicillin, and aztreonam. Average positive predictive values for genotypic prediction of phenotypic resistance were 91% for E. coli, 93% for K. pneumoniae, 87% for P. mirabilis, and 92% for P. aeruginosa across the various antibiotics for this highly resistant cohort of bacterial isolates.

SNP-IT Tool for Identifying Subspecies and Associated Lineages of Mycobacterium tuberculosis Complex

The clinical phenotype of zoonotic tuberculosis and its contribution to the global burden of disease are poorly understood and probably underestimated. This shortcoming is partly because of the inability of currently available laboratory and in silico tools to accurately identify all subspecies of the Mycobacterium tuberculosis complex (MTBC). We present SNPs to Identify TB (SNP-IT), a single-nucleotide polymorphism-based tool to identify all members of MTBC, including animal clades. By applying SNP-IT to a collection of clinical genomes from a UK reference laboratory, we detected an unexpectedly high number of M. orygis isolates. M. orygis is seen at a similar rate to M. bovis, yet M. orygis cases have not been previously described in the United Kingdom. From an international perspective, it is possible that M. orygis is an underestimated zoonosis. Accurate identification will enable study of the clinical phenotype, host range, and transmission mechanisms of all subspecies of MTBC in greater detail.

Overcoming the pitfalls of automatic interpretation of whole genome sequencing data by online tools for the prediction of pyrazinamide resistance in Mycobacterium tuberculosis

OBJECTIVES

Automated online software tools that analyse whole genome sequencing (WGS) data without the need for bioinformatics expertise can motivate the implementation of WGS-based molecular drug susceptibility testing (DST) in routine diagnostic settings for tuberculosis (TB). Pyrazinamide (PZA) is a key drug for current and future TB treatment regimens; however, it was reported that predictive power for PZA resistance by the available tools is low. Therefore, this low predictive power may make users hesitant to use the tools. This study aimed to elucidate why and to uncover the real performance of the tools when taking into account their variation calling lists (manual inspection), not just their automated reporting system (default setting) that was evaluated by previous studies.

METHODS

WGS data from 191 datasets comprising 108 PZA-resistant and 83 susceptible strains were used to evaluate the potential performance of the available online tools (TB Profiler, TGS-TB, PhyResSE, and CASTB) for predicting phenotypic PZA resistance.

RESULTS

When taking into consideration the variation calling lists, 73 variants in total (47 non-synonymous mutations and 26 indels) in pncA were detected by TGS-TB and PhyResSE, covering all mutations for the 108 PZA-resistant strains. The 73 variants were confirmed by Sanger sequencing. TB Profiler also detected all but three complete loss, two large deletion at the 3'-end, and one relatively large insertion of pncA. On the other hand, many of the 73 variants were lacking in the automated reporting systems except by TGS-TB; of these variants, CASTB detected only 20. By applying the 'non-wild type sequence' approach for predicting PZA resistance, accuracy of the results significantly improved compared with that of the automated results obtained by each tool.

CONCLUSION

Users can obtain more accurate predictions for PZA resistance than previously reported by manually checking the results and applying the 'non-wild type sequence' approach.

Genomic prediction of tuberculosis drug-resistance: benchmarking existing databases and prediction algorithms

Background

It is possible to predict whether a tuberculosis (TB) patient will fail to respond to specific antibiotics by sequencing the genome of the infecting Mycobacterium tuberculosis (Mtb) and observing whether the pathogen carries specific mutations at drug-resistance sites. This advancement has led to the collation of TB databases such as PATRIC and ReSeqTB that possess both whole genome sequences and drug resistance phenotypes of infecting Mtb isolates. Bioinformatics tools have also been developed to predict drug resistance from whole genome sequencing (WGS) data.

Here, we evaluate the performance of four popular tools (TBProfiler, MyKrobe, KvarQ, PhyResSE) with 6746 isolates compiled from publicly available databases, and subsequently identify highly probable phenotyping errors in the databases by genetically predicting the drug phenotypes using all four software.

Results

Our results show that these bioinformatics tools generally perform well in predicting the resistance status for two key first-line agents (isoniazid, rifampicin), but the accuracy is lower for second-line injectables and fluoroquinolones. The error rates in the databases are also non-trivial, reaching as high as 31.1% for prothionamide, and that phenotypes from ReSeqTB are more susceptible to errors.

Conclusions

The good performance of the automated software for drug resistance prediction from TB WGS data shown in this study further substantiates the usefulness and promise of utilising genetic data to accurately profile TB drug resistance, thereby reducing misdiagnoses arising from error-prone culture-based drug susceptibility testing.

Electronic supplementary material

The online version of this article (10.1186/s12859-019-2658-z) contains supplementary material, which is available to authorized users.

Mycobacterium tuberculosis Next-Generation Whole Genome Sequencing: Opportunities and Challenges

Mycobacterium tuberculosis drug resistance is a threat to global tuberculosis (TB) control. Comprehensive and timely drug susceptibility determination is critical to inform appropriate treatment of drug-resistant tuberculosis (DR-TB). Phenotypic drug susceptibility testing (DST) is the gold standard for M. tuberculosis drug resistance determination. M. tuberculosis whole genome sequencing (WGS) has the potential to be a one-stop method for both comprehensive DST and epidemiological investigations. We discuss in this review the tremendous opportunities that next-generation WGS presents in terms of understanding the molecular epidemiology of tuberculosis and mechanisms of drug resistance. The potential clinical value and public health impact in the areas of DST for patient management and tracing of transmission chains for timely public health intervention are also discussed. We present the current challenges for the implementation of WGS in low and middle-income settings. WGS analysis has already been adapted routinely in laboratories to inform patient management and public health interventions in low burden high-income settings such as the United Kingdom. We predict that the technology will be adapted similarly in high burden settings where the impact on the epidemic will be greatest.