Machine learning for classifying tuberculosis drug-resistance from DNA sequencing data

Yang, Yang; Niehaus, Katherine E.; Walker, Timothy M.; Iqbal, Zamin; Walker, A Sarah; Wilson, Daniel J.; Peto, Tim E. A.; Crook, Derrick W.; Smith, E. Grace; Zhu, Tingting; Clifton, David A.

doi:10.1093/bioinformatics/btx801

Cited by 121 publications

(138 citation statements)

References 16 publications

Supporting

Mentioning

132

Contrasting

Unclassified

Order By: Relevance

“…Improvement in portable sequencing technology 60 and decreased cost of sequencing is promising to facilitate adoption in settings with lower resources were TB is most prevalent. However even if sequencing technology is available, our results suggest that genomic data interpretation will likely necessitate the use of statistical models or machine learning 16,58,61 given the number of genetic loci associated with resistance and the likely contribution of gene-gene interactions, especially if a quantitative prediction of the drug MIC is desirable 15 . The portability of the potential benefits of these advances to areas of the world where TB is most prevalent will require continued efforts in open sharing of data and analysis tools 62 .…”

Section: Para-aminosalicylic Acid (Pas)mentioning

confidence: 99%

Genome wide association with quantitative resistance phenotypes in Mycobacterium tuberculosis reveals novel resistance genes and regulatory regions

Mustafa

Freschi

Calderón

et al. 2018

Preprint

View full text Add to dashboard Cite

Drug resistance is threatening attempts at tuberculosis epidemic control. Molecular diagnostics for drug resistance that rely on the detection of resistance-related mutations could expedite patient care and accelerate progress in TB eradication. We performed minimum inhibitory concentration testing for 12 anti-TB drugs together with Illumina whole genome sequencing on 1452 clinical Mycobacterium tuberculosis (MTB) isolates. We then used a linear mixed model to evaluate genome wide associations between mutations in MTB genes or noncoding regions and drug resistance, followed by validation of our findings in an independent dataset of 792 patient isolates. Novel associations at 13 genomic loci were confirmed in the validation set, with 2 involving noncoding regions. We found promoter mutations to have smaller average effects on resistance levels than gene body mutations in genes where both can contribute to resistance. Enabled by a quantitative measure of resistance, we estimated the heritability of the resistance phenotype to 11 anti-TB drugs and identify a lower than expected contribution from known resistance genes. We also report the proportion of variation in resistance levels explained by the novel loci identified here. This study highlights the complexity of the genomic mechanisms associated with the MTB resistance phenotype, including the relatively large number of potentially causative or compensatory loci, and emphasizes the contribution of the noncoding portion of the genome.2 Introduction: Tuberculosis (TB) remains a major global public health threat. In 2016 there were an estimated 10.4 million TB cases globally and 1.7 million deaths due to the disease. One of the most challenging forms of disease is caused by multidrug resistant (MDR) Mycobacterium tuberculosis, with a global annual incidence of over half a million cases 1 . The World Health Organization (WHO) estimates that only two of every three patients with multidrug resistant TB are diagnosed, three in every four of the diagnosed are treated, and only one of every two of the treated patients are cured, resulting in the grim reality of about 75% of the incident cases persisting in the community or succumbing to their illness. Antibiotic resistance is also an increasing problem in other human pathogens, and transmission of antibiotic resistance from person to person is amplifying the public health threat 2 .Improved surveillance, diagnosis and treatment are designated priorities by the WHO and the US, European CDCs for addressing the antibiotic resistance challenge 1,3,4 . These measures will rely on an improved understanding of the mechanisms of resistance acquisition in bacteria. The knowledge of genetic mechanisms of antibiotic resistance has formed the basis of several commercial molecular diagnostics for TB that have had remarkable global uptake, despite the fact that they only reliably test for a subset of TB drugs and hence have not yet been able to replace the traditional more costly and slow process of mycobacterial culture and drug susce...

show abstract

Section: Para-aminosalicylic Acid (Pas)mentioning

confidence: 99%

Genome wide association with quantitative resistance phenotypes in Mycobacterium tuberculosis reveals novel resistance genes and regulatory regions

Mustafa

Freschi

Calderón

et al. 2018

Preprint

View full text Add to dashboard Cite

show abstract

“…Past studies utilizing whole genome sequencing have shown a wide range of performance, with sensitivities for first-line drugs ranging from 54% to 98% [8,12,13]. Second-line injectable drugs and fluoroquinolones had lower sensitivities, most of which were between 30% and 96% [8,12,13]. We hypothesize that the limited predictive performance of anti-tuberculosis drugs outside of first-line drugs could be improved using a large dataset enriched for resistance to second-line drugs and a more complex model.Deep learning models have become a powerful tool for many classification tasks.…”

mentioning

confidence: 93%

“…The limited scope of these tests suggests the need for a comprehensive antimicrobial susceptibility test.An alternative to targeted mutation detection methods is whole genome sequencing, which captures both common and rare mutations involved in drug resistance. Past studies utilizing whole genome sequencing have shown a wide range of performance, with sensitivities for first-line drugs ranging from 54% to 98% [8,12,13]. Second-line injectable drugs and fluoroquinolones had lower sensitivities, most of which were between 30% and 96% [8,12,13].…”

mentioning

confidence: 99%

Deep learning predicts tuberculosis drug resistance status from genome sequencing data

Chen

Doddi

Royer

et al. 2018

Preprint

View full text Add to dashboard Cite

One sentence summary: A unified multitask deep learning model can be used to identify multidrug resistant Mycobacterium tuberculosis using sequencing data. AbstractThe diagnosis of multidrug resistant and extensively drug resistant tuberculosis is a global health priority. Whole genome sequencing of clinical Mycobacterium tuberculosis isolates promises to circumvent the long wait times and limited scope of conventional phenotypic drug susceptibility but gaps remain for predicting phenotype accurately from genotypic data. Using targeted or whole genome sequencing and conventional drug resistance phenotyping data from 3,601 Mycobacterium tuberculosis strains, 1,228 of which were multidrug resistant, we implemented the first multitask deep learning framework to predict phenotypic drug resistance to 10 anti-tubercular drugs. The proposed wide and deep neural network (WDNN) achieved improved predictive performance compared to regularized logistic regression and random forest: the average sensitivities and specificities, respectively, were 92.7% and 92.7% for first-line drugs and 82.0% and 92.8% for second-line drugs during cross-validation. On an independent validation set, the multitask WDNN showed significant performance gains over baseline models, with average sensitivities and specificities, respectively, of 84.5% and 93.6% for first-line drugs and 64.0% and 95.7% for second-line drugs. In addition to being able to learn from samples that have only been partially phenotyped, our proposed multitask architecture shares information across different anti-tubercular drugs and genes to provide a more accurate phenotypic prediction. We use t-distributed Stochastic Neighbor Embedding (t-SNE) visualization and feature importance analyses to examine inter-drug similarities. Deep learning has a clear role in improving drug resistance predictive performance over traditional methods and holds promise in bringing sequencing technologies closer to the bedside.. CC-BY-NC-ND 4.0 International license It is made available under a was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. Diagnosing drug resistance remains a barrier to providing appropriate TB treatment. Due to insufficient resources for building diagnostic laboratories, fewer than half of the countries with a high MDR-TB burden have modern diagnostic capabilities (3). Even in the best equipped laboratories, conventional culture and culture based drug susceptibility testing (DST) constitutes a considerable biohazard and requires weeks to months before results are reported due to Mycobacterium tuberculosis's slow growth in vitro (1). Molecular diagnostics are now an increasingly common alternative to conventional cultures. The WHO has endorsed three such molecular tests: the GeneXpert MTB/RIF a rapid RT-PCR based diagnostic test assay that detects RIF resistance, the Hain line probe assay (LPA) that tests for both RIF and INH resistance, and the Hain MDRTBsl an LPA that tests for resistance to second-line in...

show abstract

“…To date, these methods have been restricted mostly to assign bacteria to 74 binary categories, i.e. susceptible or non-susceptible (12), (13), (14), (8), (15), (16), (17). 75 However, clinical breakpoints used to define susceptible and non-susceptible categories 76 can change and such binary categories do not allow following more subtle changes in 77 susceptibility in time.…”

Section: Introduction 48mentioning

confidence: 99%

Understanding and predicting ciprofloxacin minimum inhibitory concentration in Escherichia coli with machine learning

Pataki

Matamoros

Putten

et al. 2019

Preprint

View full text Add to dashboard Cite

A possible way to slow down the antibiotic resistance crisis is to be 17 more strict when it comes to antibiotics prescriptions. For accurate antibiotic pre-18 scriptions, antibiotic susceptibility data are needed. With the increasing availability 19 of next-generation sequencing (NGS), bacterial whole genome sequencing (WGS) is 20 becoming a feasible alternative to traditional phenotyping for the detection and surveil-21 lance of AMR. 22 Pataki et al. the flexibility to follow potential adjustments in definitions of susceptible and resistant 44 categories (breakpoints). 45 KEYWORDS: AMR, MIC, machine learning, antibiotics, personalized medicine, 46 ciprofloxacin 47 102 RESULTS 103 Data 704 E. coli genomes were analyzed in this study which had MIC measurement 104 for ciprofloxacin (24). Paired-end sequencing was performed on all of them and the 105 results were stored in FASTQ format. The samples originated from Denmark, Italy, USA, 106 UK, and Vietnam. 266 out of the 704 E. coli genomes had no country metadata available

show abstract

Machine learning for classifying tuberculosis drug-resistance from DNA sequencing data

Cited by 121 publications

References 16 publications

Genome wide association with quantitative resistance phenotypes in Mycobacterium tuberculosis reveals novel resistance genes and regulatory regions

Genome wide association with quantitative resistance phenotypes in Mycobacterium tuberculosis reveals novel resistance genes and regulatory regions

Deep learning predicts tuberculosis drug resistance status from genome sequencing data

Understanding and predicting ciprofloxacin minimum inhibitory concentration in Escherichia coli with machine learning

Contact Info

Product

Resources

About