Whole genome sequencing of hepatitis B virus using tiled amplicon (HEPTILE) and probe based enrichment on Illumina and Nanopore platforms

Abstract

Hepatitis B virus (HBV) whole genome sequencing (WGS) is currently limited as the DNA viral loads (VL) of many clinical samples are below the threshold required to generate full genomes using current sequencing methods. We developed two pan-genotypic viral enrichment methods, using probe-based capture and tiled amplicon PCR (HEP-TILE) for HBV WGS. We demonstrate using mock samples that both enrichment methods are pan-genotypic (genotypes A-J). Using clinical samples, we demonstrate that HEP-TILE amplification successfully amplifies full genomes at the lowest HBV VL tested (30 IU/ml), and the PCR products can be sequenced using both Nanopore and Illumina platforms. Probe-based capture with Illumina sequencing required VL > 300,000 IU/ml to generate full length HBV genomes. The capture-Illumina and HEP-TILE-Nanopore pipelines had consensus sequencing accuracy of 100% in mock samples with known DNA sequences. Together, these protocols will facilitate the generation of HBV sequence data, enabling a more accurate and representative picture of HBV molecular epidemiology, cast light on persistence and pathogenesis, and enhance understanding of the outcomes of infection and its treatment.

Keywords: Hepatitis B virus (HBV); Illumina; Nanopore; Whole genome sequencing.

Fig. 1

Design of HBV targeted enrichment methods. (A) HBV genome structure: relaxed circular DNA (rc-DNA). The complete negative strand is discontinuous, with a ‘nick’ and covalently bound polymerase. The positive strand is incomplete with a variable length and RNA tail. (B) Probe design schematic for HBV probe based enrichment. An initial set of probes (red) are designed based on non-recombinant whole genomes downloaded from HBVdb. Additional probes (orange) are added in regions of high diversity, using an extended reference sequence to account for the circular genome, and to account for insertions/deletions, then the final probe panel confirmed. Blue dots indicate 5’ biotinylation. (C) Multiplex tiling PCR for the circular HBV genome. Generation of tiled schemes from linear reference genomes leaves a gap in coverage for a circular genome. An additional amplicon spanning the start/end of the linear reference ensures a theoretical maximum 100% genome coverage can be obtained. Primer positions and amplicon length for the final HEP-TILE scheme (hbv/600/v2.1.0). Numbering of nt positions based on X02763 reference genome. Created in BioRender. Lumley, S. (2024) BioRender.com/r40b439.

Fig. 2

Overview of sequencing workflows. Numbers of samples sequenced at each stage indicated. Due to sample availability, project stage, and local infrastructure, we used different subsets of samples to validate each method. In the clinical sample set from South Africa, if residual extract was available after capture, two further aliquots were taken forward for HEP-TILE enrichment and sequencing on Nanopore and Illumina. Created in BioRender. Lumley, S. (2024) BioRender.com/k41p399.

Fig. 3

HBV genomes for all HBV genotypes (A–J) in mock clinical samples enriched using HBV probes and primers (A) Enriched with probe based capture, sequenced on Illumina; (B) Enriched with HEP-TILE, sequenced on Nanopore; (C) Enriched with HEP-TILE, sequenced on Illumina.

Fig. 4

HBV WGS using capture and amplicon based target enrichment. Top row: Relationship between viral load and percentage of reads on-target (i.e. HBV) for (A) probe-based capture sequenced on Illumina, (B) HEP-TILE sequenced on Nanopore and (C) HEP-TILE sequenced on Illumina. Bottom row: Relationship between viral load and percentage of genome with high confidence consensus for (D) probe-based capture sequenced on Illumina (x5 minimum depth), (E) HEP-TILE sequenced on Nanopore (x20 minimum depth) and (F) HEP-TILE sequenced on Illumina (x5 minimum depth). Points are coloured by the percentage genome coverage. Notes. Panels E and F show an outlier with viral load 5.41 log₁₀ IU/ml, with disproportionately 0% genome coverage. Repeat VL was 3 log₁₀ IU/ml. There is one sample, HEP-1529, in the lower left corner of Fig. 4B with 0.55% reads (260 reads) mapped to the HBV genome, which led to 45% coverage as shown in supplementary Table 2. This is mathematically possible as the minimum theoretical number of reads to obtain a whole genome at x20 depth in a 6 amplicon scheme is only 120 reads.