Nanopore sequencing enhances the detection of low-frequency variants and mutational linkage in HIV-1 drug resistance.

OBJECTIVES: Low-frequency HIV-1 drug resistance mutations (DRMs) and mutation linkage within viral genomes may contribute to treatment failure and multidrug resistance but are difficult to resolve using conventional sequencing. We evaluated the performance of nanopore long-read sequencing for HIV-1 DRM detection and haplotype analysis. METHODS: In this retrospective study, 230 plasma samples from people living with HIV undergoing resistance testing at Beijing Ditan Hospital between Aug 2024 and July 2025 were analysed using the nanopore-based G-seq500 platform. DRM profiles and subtype classifications were compared with Sanger sequencing, and discordant mutations were validated by next-generation sequencing (NGS, DNBSEQ-T7 platform). Long-read data were further used for haplotype reconstruction. RESULTS: The overall concordance rate between G-seq500 and Sanger sequencing for DRM detection was 93.0%. Additional resistance-associated mutations identified by G-seq500 were confirmed by NGS and were predominantly low-frequency variants. Subtyping concordance was 92.2% (212/230). Haplotype reconstruction revealed substantial intra-host viral diversity, with nearly half of samples containing multiple viral haplotypes. Multiple DRMs frequently co-occurred within the same haplotype, indicating linked multidrug-resistant viral populations. CONCLUSIONS: G-seq500 sequencing showed comparable performance to Sanger sequencing for HIV-1 DRM detection and subtype classification. Long-read sequencing further enabled characterization of low-frequency resistance-associated variants and within-host viral haplotype structures, providing additional insights into the complexity of HIV-1 drug resistance beyond conventional consensus sequencing.