Rapid Whole-Genome Sequencing of Mycobacterium tuberculosis Isolates Directly from Clinical Samples

Abstract

The rapid identification of antimicrobial resistance is essential for effective treatment of highly resistant Mycobacterium tuberculosis. Whole-genome sequencing provides comprehensive data on resistance mutations and strain typing for monitoring transmission, but unlike for conventional molecular tests, this has previously been achievable only from cultures of M. tuberculosis. Here we describe a method utilizing biotinylated RNA baits designed specifically for M. tuberculosis DNA to capture full M. tuberculosis genomes directly from infected sputum samples, allowing whole-genome sequencing without the requirement of culture. This was carried out on 24 smear-positive sputum samples, collected from the United Kingdom and Lithuania where a matched culture sample was available, and 2 samples that had failed to grow in culture. M. tuberculosis sequencing data were obtained directly from all 24 smear-positive culture-positive sputa, of which 20 were of high quality (>20× depth and >90% of the genome covered). Results were compared with those of conventional molecular and culture-based methods, and high levels of concordance between phenotypical resistance and predicted resistance based on genotype were observed. High-quality sequence data were obtained from one smear-positive culture-negative case. This study demonstrated for the first time the successful and accurate sequencing of M. tuberculosis genomes directly from uncultured sputa. Identification of known resistance mutations within a week of sample receipt offers the prospect for personalized rather than empirical treatment of drug-resistant tuberculosis, including the use of antimicrobial-sparing regimens, leading to improved outcomes.

FIG 1

Mean coverage and percentage of on-target reads (OTR) in sequencing from sputum with and without enrichment for two samples.

FIG 2

(A) Depth of coverage obtained for smear-positive samples from sputum and culture. (B) Depth of coverage for sputum sequence from smear-positive samples which failed to grow. The level of smear positivity is shown, with the remaining samples being smear negative.

FIG 3

Resistance phenotype and genotype of matched pairs. R, a mutation exists at a level of greater than 10%. Low R, a mutation in codon 306 of the embB gene which is thought to confer low-level resistance to ethambutol. NA, not applicable, as phenotype was not tested. Rif, rifampin; Inh, isoniazid; Emb, ethambutol; Pza, pyrazinamide; Str, streptomycin; Ofl, ofloxacin (fluoroquinolones); Pas, para-aminosalicylic acid; Amg, aminoglycosides; Thi, thionamides. The asterisks indicate second-line antibiotics.

FIG 4

Heteroresistance in gyrA in patient 10. R, resistant allele with suffix indicating codon position; S, absence of resistant allele.