Use of whole-genome sequencing to predict Mycobacterium tuberculosis drug resistance in Shanghai, China

Objective To evaluate the performance of whole-genome sequencing (WGS) for predicting Mycobacterium tuberculosis (MTB) drug resistance. Methods 276 rifampin-resistance tuberculosis (RR-TB) and 30 rifampicin-sensitive clinical isolates were randomly selected from patients with tuberculosis in Shanghai Pulmonary Hospital (SPH). Phenotypic drug susceptibility testing (DST) against six anti-TB drugs was performed, and WGS was used to predict the drug resistance using an online 'TB-Profiler' tool. Results Using phenotypic susceptibility as the gold standard, the overall sensitivities and specificities for WGS were 94.53% and 92.00% for isoniazid, 97.10% and 100.00% for rifampicin, 97.46% and 64.36% for ethambutol, 97.14% and 95.83% or streptomycin, 93.02% and 98.87% for ofloxacin, and 75.00% and 100.00% for amikacin, respectively. The concordances of WGS-based DST and phenotypic DST were: isoniazid (94.12%), rifampicin (97.39%), ethambutol (77.12%), streptomycin (96.73%), ofloxacin (96.41%) and amikacin (97.06%). Conclusions WGS could be a promising approach to predict resistance to isoniazid, rifampicin, ethambutol, streptomycin, ofloxacin, and amikacin.