{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,9,12]],"date-time":"2026-09-12T11:30:05Z","timestamp":1789212605304,"version":"build-2803163510"},"update-to":[{"DOI":"10.1371\/journal.pcbi.1012260","type":"new_version","label":"New version","source":"publisher","updated":{"date-parts":[[2024,8,15]],"date-time":"2024-08-15T00:00:00Z","timestamp":1723680000000}}],"reference-count":31,"publisher":"Public Library of Science (PLoS)","issue":"8","license":[{"start":{"date-parts":[[2024,8,5]],"date-time":"2024-08-05T00:00:00Z","timestamp":1722816000000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"name":"See manuscript for full list"}],"content-domain":{"domain":["www.ploscompbiol.org"],"crossmark-restriction":false},"short-container-title":["PLoS Comput Biol"],"abstract":"<jats:p>There remains a clinical need for better approaches to rapid drug susceptibility testing in view of the increasing burden of multidrug resistant tuberculosis. Binary susceptibility phenotypes only capture changes in minimum inhibitory concentration when these cross the critical concentration, even though other changes may be clinically relevant. We developed a machine learning system to predict minimum inhibitory concentration from unassembled whole-genome sequencing data for 13 anti-tuberculosis drugs. We trained, validated and tested the system on 10,859 isolates from the CRyPTIC dataset. Essential agreement rates (predicted MIC within one doubling dilution of observed MIC) were above 92% for first-line drugs, 91% for fluoroquinolones and aminoglycosides, and 90% for new and repurposed drugs, albeit with a significant drop in performance for the very few phenotypically resistant isolates in the latter group. To further validate the model in the absence of external MIC datasets, we predicted MIC and converted values to binary for an external set of 15,239 isolates with binary phenotypes, and compare their performance against a previously validated mutation catalogue, the expected performance of existing molecular assays, and World Health Organization Target Product Profiles. The sensitivity of the model on the external dataset was greater than 90% for all drugs except ethionamide, clofazimine and linezolid. Specificity was greater than 95% for all drugs except ethambutol, ethionamide, bedaquiline, delamanid and clofazimine. The proposed system can provide quantitative susceptibility phenotyping to help guide antimicrobial therapy, although further data collection and validation are required before machine learning can be used clinically for all drugs.<\/jats:p>","DOI":"10.1371\/journal.pcbi.1012260","type":"journal-article","created":{"date-parts":[[2024,8,5]],"date-time":"2024-08-05T13:44:39Z","timestamp":1722865479000},"page":"e1012260","update-policy":"https:\/\/doi.org\/10.1371\/journal.pcbi.corrections_policy","source":"Crossref","is-referenced-by-count":12,"title":["Quantitative drug susceptibility testing for Mycobacterium tuberculosis using unassembled sequencing data and machine learning"],"prefix":"10.1371","volume":"20","author":[{"name":"The CRyPTIC consortium","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"340","published-online":{"date-parts":[[2024,8,5]]},"reference":[{"key":"pcbi.1012260.ref001","unstructured":"World Health Organization. Global tuberculosis report 2021. (2021)."},{"key":"pcbi.1012260.ref002","author":"World Health Organization","year":"2021","journal-title":"Impact of the COVID-19 pandemic on TB detection and mortality in 2020."},{"key":"pcbi.1012260.ref003","author":"World Health Organization","year":"2015","journal-title":"The end TB strategy"},{"key":"pcbi.1012260.ref004","article-title":"Rapid, comprehensive, and affordable mycobacterial diagnosis with whole-genome sequencing: a prospective study","author":"J. Pankhurst","year":"2016","journal-title":"Lancet Respir. Med"},{"key":"pcbi.1012260.ref005","first-page":"e02314","article-title":"Xpert MTB\/XDR: a 10-Color Reflex Assay Suitable for Point-of-Care Settings To Detect Isoniazid, Fluoroquinolone, and Second-Line-Injectable-Drug Resistance Directly from Mycobacterium tuberculosis-Positive Sputum","volume":"59","author":"Y. Cao","year":"2021","journal-title":"J. Clin. Microbiol"},{"key":"pcbi.1012260.ref006","first-page":"e01480","article-title":"Evaluation of Whole-Genome Sequencing for Mycobacterial Species Identification and Drug Susceptibility Testing in a Clinical Setting: a Large-Scale Prospective Assessment of Performance against Line Probe Assays and Phenotyping","volume":"56","author":"T. P. Quan","year":"2018","journal-title":"J. Clin. Microbiol"},{"key":"pcbi.1012260.ref007","doi-asserted-by":"crossref","first-page":"922","DOI":"10.3389\/fgene.2019.00922","article-title":"Machine Learning Predicts Accurately Mycobacterium tuberculosis Drug Resistance From Whole Genome Sequencing Data","volume":"10","author":"W. Deelder","year":"2019","journal-title":"Front. Genet"},{"key":"pcbi.1012260.ref008","doi-asserted-by":"crossref","first-page":"5487","DOI":"10.1038\/s41598-020-62368-2","article-title":"Artificial Intelligence and Machine learning based prediction of resistant and susceptible mutations in Mycobacterium tuberculosis","volume":"10","author":"S. Jamal","year":"2020","journal-title":"Sci. Rep"},{"key":"pcbi.1012260.ref009","doi-asserted-by":"crossref","first-page":"2427","DOI":"10.1038\/s41598-022-06449-4","article-title":"Accurate and rapid prediction of tuberculosis drug resistance from genome sequence data using traditional machine learning algorithms and CNN","volume":"12","author":"X. Kuang","year":"2022","journal-title":"Sci. Rep"},{"key":"pcbi.1012260.ref010","doi-asserted-by":"crossref","first-page":"e01260","DOI":"10.1128\/JCM.01260-18","article-title":"Using Machine Learning To Predict Antimicrobial MICs and Associated Genomic Features for Nontyphoidal Salmonella","volume":"57","author":"M. Nguyen","year":"2019","journal-title":"J. Clin. Microbiol"},{"key":"pcbi.1012260.ref011","doi-asserted-by":"crossref","first-page":"3687","DOI":"10.1016\/j.sjbs.2022.02.047","article-title":"Prediction of antimicrobial minimal inhibitory concentrations for Neisseria gonorrhoeae using machine learning models","volume":"29","author":"M. Yasir","year":"2022","journal-title":"Saudi J. Biol. Sci"},{"key":"pcbi.1012260.ref012","doi-asserted-by":"crossref","first-page":"841289","DOI":"10.3389\/fmicb.2022.841289","article-title":"A Practical Approach for Predicting Antimicrobial Phenotype Resistance in Staphylococcus aureus Through Machine Learning Analysis of Genome Data.","volume":"13","author":"S. Wang","year":"2022","journal-title":"Front. Microbiol."},{"key":"pcbi.1012260.ref013","doi-asserted-by":"crossref","first-page":"15026","DOI":"10.1038\/s41598-020-71693-5","article-title":"Understanding and predicting ciprofloxacin minimum inhibitory concentration in Escherichia coli with machine learning.","volume":"10","author":"B. \u00c1. Pataki","year":"2020","journal-title":"Sci. Rep."},{"issue":"8","key":"pcbi.1012260.ref014","doi-asserted-by":"crossref","first-page":"e3001721","DOI":"10.1371\/journal.pbio.3001721","article-title":"A data compendium associating the genomes of 12,289 Mycobacterium tuberculosis isolates with quantitative resistance phenotypes to 13 antibiotics","volume":"20","author":"CRyPTIC","year":"2022","journal-title":"Plos Biology"},{"key":"pcbi.1012260.ref015","doi-asserted-by":"crossref","first-page":"2200239","DOI":"10.1183\/13993003.00239-2022","article-title":"Epidemiological cut-off values for a 96-well broth microdilution plate for high-throughput research antibiotic susceptibility testing of M. tuberculosis","volume":"60","author":"CRyPTIC","year":"2022","journal-title":"Eur. Respir. J"},{"key":"pcbi.1012260.ref016","first-page":"S2666","article-title":"The 2021 WHO catalogue of Mycobacterium tuberculosis complex mutations associated with drug resistance: a genotypic analysis","volume":"21","author":"T. Walker","year":"2022","journal-title":"The Lancet Microbe"},{"key":"pcbi.1012260.ref017","doi-asserted-by":"crossref","first-page":"1403","DOI":"10.1056\/NEJMoa1800474","article-title":"Prediction of Susceptibility to First-Line Tuberculosis Drugs by DNA Sequencing","volume":"379","author":"CRyPTIC Consortium and the 100,000 Genomes Project","year":"2018","journal-title":"N. Engl. J. Med"},{"key":"pcbi.1012260.ref018","doi-asserted-by":"crossref","first-page":"S126","DOI":"10.1093\/infdis\/jiz106","article-title":"Guidance for Studies Evaluating the Accuracy of Rapid Tuberculosis Drug-Susceptibility Tests","volume":"220","author":"S. Georghiou","year":"2019","journal-title":"J. Infect. Dis"},{"key":"pcbi.1012260.ref019","article-title":"WHO consolidated guidelines on tuberculosis: module 4: treatment: drug-resistant tuberculosis treatment","author":"World Health Organization","year":"2020"},{"key":"pcbi.1012260.ref020","doi-asserted-by":"crossref","first-page":"e01677","DOI":"10.1128\/JCM.01677-19","article-title":"Validation of Bedaquiline Phenotypic Drug Susceptibility Testing Methods and Breakpoints: a Multilaboratory, Multicountry Study","volume":"58","author":"K. Kaniga","year":"2020","journal-title":"J. Clin. Microbiol"},{"key":"pcbi.1012260.ref021","doi-asserted-by":"crossref","first-page":"2001796","DOI":"10.1183\/13993003.01796-2020","article-title":"Rapid genomic first- and second-line drug resistance prediction from clinical Mycobacterium tuberculosis specimens using Deeplex-MycTB","volume":"57","author":"S. Feuerriegel","year":"2021","journal-title":"Eur. Respir. J"},{"key":"pcbi.1012260.ref022","article-title":"Target product profile for next-generation drug-susceptibility testing at peripheral centres.","author":"World Health Organization","year":"2021"},{"key":"pcbi.1012260.ref023","doi-asserted-by":"crossref","first-page":"51","DOI":"10.1038\/s41579-018-0098-9","article-title":"Developmental roadmap for antimicrobial susceptibility testing systems","volume":"17","author":"A. van Belkum","year":"2019","journal-title":"Nat. Rev. Microbiol"},{"key":"pcbi.1012260.ref024","article-title":"Artificial Intelligence and Machine Learning in Software as a Medical Device.","author":"Food and Drug Administration","year":"2021"},{"key":"pcbi.1012260.ref025","doi-asserted-by":"crossref","first-page":"118","DOI":"10.1038\/s41746-020-00324-0","article-title":"The state of artificial intelligence-based FDA-approved medical devices and algorithms: an online database.","volume":"3","author":"S. Benjamens","year":"2020","journal-title":"NPJ Digit. Med"},{"key":"pcbi.1012260.ref026","article-title":"Genomic variant-identification methods may alter Mycobacterium tuberculosis transmission inferences","author":"K. S. Walter","year":"2020","journal-title":"Microb. Genomics"},{"key":"pcbi.1012260.ref027","first-page":"27930","volume":"6","author":"J. J. Davis","year":"2016","journal-title":"Antimicrobial Resistance Prediction in PATRIC and RAST. Sci. Rep"},{"key":"pcbi.1012260.ref028","doi-asserted-by":"crossref","first-page":"754","DOI":"10.1186\/s12864-016-2889-6","article-title":"Predictive computational phenotyping and biomarker discovery using reference-free genome comparisons","volume":"17","author":"A. Drouin","year":"2016","journal-title":"BMC Genomics"},{"key":"pcbi.1012260.ref029","doi-asserted-by":"crossref","first-page":"16041","DOI":"10.1038\/nmicrobiol.2016.41","article-title":"Identifying lineage effects when controlling for population structure improves power in bacterial association studies","volume":"1","author":"S. G. Earle","year":"2016","journal-title":"Nat. Microbiol"},{"key":"pcbi.1012260.ref030","doi-asserted-by":"crossref","first-page":"62","DOI":"10.1128\/AAC.50.1.62-67.2006","article-title":"Comparison of censored regression and standard regression analyses for modeling relationships between antimicrobial susceptibility and patient- and institution-specific variables","volume":"50","author":"J. P. Hammel","year":"2006","journal-title":"Antimicrob. Agents Chemother"},{"key":"pcbi.1012260.ref031","doi-asserted-by":"crossref","unstructured":"Chen, T. & Guestrin, C. XGBoost: A Scalable Tree Boosting System. in Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining 785\u2013794 (ACM, 2016). doi: 10.1145\/2939672.2939785","DOI":"10.1145\/2939672.2939785"}],"updated-by":[{"DOI":"10.1371\/journal.pcbi.1012260","type":"new_version","label":"New version","source":"publisher","updated":{"date-parts":[[2024,8,15]],"date-time":"2024-08-15T00:00:00Z","timestamp":1723680000000}}],"container-title":["PLOS Computational Biology"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/dx.plos.org\/10.1371\/journal.pcbi.1012260","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2024,8,15]],"date-time":"2024-08-15T13:56:33Z","timestamp":1723730193000},"score":1,"resource":{"primary":{"URL":"https:\/\/dx.plos.org\/10.1371\/journal.pcbi.1012260"}},"subtitle":[],"editor":[{"given":"Luigi","family":"Palla","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"editor"}]}],"short-title":[],"issued":{"date-parts":[[2024,8,5]]},"references-count":31,"journal-issue":{"issue":"8","published-online":{"date-parts":[[2024,8,5]]}},"URL":"https:\/\/doi.org\/10.1371\/journal.pcbi.1012260","relation":{},"ISSN":["1553-7358"],"issn-type":[{"value":"1553-7358","type":"electronic"}],"subject":[],"published":{"date-parts":[[2024,8,5]]}}}