Application of next-generation sequencing to detect variants of drug-resistant Mycobacterium tuberculosis: genotype–phenotype correlation

Dae-Hyun Ko,
Eun Jin Lee,
Su-Kyung Lee,
Han-Sung Kim,
So Youn Shin,
Jungwon Hyun,
Jae-Seok Kim,
Wonkeun Song and
Hyun Soo KimEmail author View ORCID ID profile

Annals of Clinical Microbiology and Antimicrobials201918:2

https://doi.org/10.1186/s12941-018-0300-y

Received: 7 August 2018
Accepted: 26 December 2018
Published: 3 January 2019

Abstract

Background

Drug resistance in Mycobacterium tuberculosis (MTB) is a major health issue worldwide. Recently, next-generation sequencing (NGS) technology has begun to be used to detect resistance genes of MTB. We aimed to assess the clinical usefulness of Ion S5 NGS TB research panel for detecting MTB resistance in Korean tuberculosis patients.

Methods

Mycobacterium tuberculosis with various drug resistance profiles including susceptible strains (N = 36) were isolated from clinical specimens. Nucleic acids were extracted from inactivated culture medium and underwent amplicon-based NGS to detect resistance variants in eight genes (gyrA, rpoB, pncA, katG, eis, rpsL, embB, and inhA). Data from previous studies using the same panel were merged to yield pooled sensitivity and specificity values for detecting drug resistance compared to phenotype-based methods.

Results

The sequencing reactions were successful for all samples. A total of 24 variants were considered to be related to resistance, and 6 of them were novel. Agreement between the phenotypic and genotypic results was excellent for isoniazid, rifampicin, and ethambutol, and was poor for streptomycin, amikacin, and kanamycin. The negative predictive values were greater than 97% for all drug classes, while the positive predictive values varied (44% to 100%). There was a possibility that common mutations could not be detected owing to the low coverage.

Conclusions

We successfully applied NGS for genetic analysis of drug resistances in MTB, as well as for susceptible strains. We obtained lists of polymorphisms and possible polymorphisms, which could be used as a guide for future tests applying NGS in mycobacteriology laboratories. When analyzing the results of NGS, coverage analysis of each samples for each gene and benign polymorphisms not related to drug resistance should be considered.