Detection of heteroresistance in Mycobacterium tuberculosis using nanopore-based amplicon sequencing and adaptive sampling

Abstract

Abstract Heteroresistance (HR) in Mycobacterium tuberculosis —the coexistence of drug-susceptible and drug-resistant subpopulations within the same host—poses a major obstacle to effective tuberculosis (TB) diagnosis, treatment, and control. The primary objective of this study was to determine whether Oxford Nanopore Technologies (ONT)-based workflows could accurately detect heteroresistance and compare their performance. To achieve this, we first optimized ONT sequencing workflows using cultured isolates and TB-positive clinical sputum, identifying phenol-based DNA extraction with ligation library as the best-performing combination, though Kit-Ligation had the best performance in amplicon sequencing from culture. We then evaluated two enrichment strategies—amplicon sequencing and Adaptive sampling (AS)—using synthetic mixtures of drug-susceptible (H37Rv) and drug-resistant (DM97) strains at defined ratios. Amplicon sequencing achieved coverage > 4,000× across target genes and reliably detected resistant alleles at frequencies as low as 1%. AS enriched 14 resistance loci with ~ 3-fold higher on-target depth compared to non-enriched nanopore whole-genome sequencing (WGS), enabling accurate detection of resistant subpopulations at 5% while retaining genomic context for lineage assignment. Together, these results demonstrate that, among the ONT workflows evaluated, nanopore amplicon sequencing provides superior sensitivity, whereas AS offers an intermediate approach, increasing depth while retaining some genomic context. These complementary approaches represent scalable strategies that could enhance TB heteroresistance detection and surveillance in high-burden settings.