Evaluation of a multiplexed tiling PCR scheme for whole-genome amplification of hepatitis B virus using Oxford Nanopore sequencing

This study evaluates a tiling PCR primer scheme adapted for Oxford Nanopore sequencing of Hepatitis B virus, finding that while the method is feasible, uneven amplification efficiency across amplicons and strong dependence on viral load (Ct values) limit consistent full-genome recovery, necessitating further optimization of the primer design.

The Gist

The Big Picture: Trying to Rebuild a Shattered Puzzle

Imagine the Hepatitis B virus (HBV) is a tiny, circular puzzle. To understand the virus, track its mutations, and see if it's becoming resistant to medicine, scientists need to see the entire puzzle, not just a few pieces.

For a long time, scientists could only look at a few specific pieces (like the "Polymerase" or "Surface" sections). But this is like trying to understand a whole movie by only watching the opening credits. You miss the plot twists and the ending.

The Goal: The researchers wanted to use a new, fast technology called Oxford Nanopore sequencing to photograph the entire viral puzzle at once.

The Tool: A "Tiling" Net

To get the whole puzzle, you can't just take one giant photo; the virus is too small and complex. Instead, scientists use a method called Tiling PCR.

Think of this like a fishing net.

• Instead of one giant net that might miss things, they use many small, overlapping nets (called amplicons).
• They cast 10 different nets across the virus's genome.
• Because the nets overlap, if one net misses a fish, the neighbor net catches it. This ensures they get the whole picture.

The researchers took a set of nets designed by another team (Ringlander et al.) that was originally built for a different type of camera (Ion Torrent). They wanted to see if they could strip off the old camera parts and use these nets with the new Nanopore camera.

The Experiment: One Big Bucket vs. Two Separate Buckets

The team tested two ways to cast these nets:

1. The "All-in-One" Bucket: They threw all 10 nets into a single bucket of water (one PCR reaction) and cast them all at once.
2. The "Split" Bucket: They divided the nets into two groups (odd-numbered nets in one bucket, even-numbered in another) and cast them separately.

The Analogy: Imagine trying to catch 10 different types of fish.

• Strategy A: Throw all 10 nets in at once. It's faster and easier, but the nets might get tangled or compete for space.
• Strategy B: Throw 5 nets, wait, then throw the other 5. It takes twice as long and doubles the work, but the nets don't tangle.

The Result: Surprisingly, it didn't matter much which strategy they used. Both methods gave similar results. The "All-in-One" bucket was just as effective as the "Split" buckets.

The Problem: The "Weak Links" in the Chain

Here is where the story gets interesting. Even though the method worked, the results were uneven.

Imagine the 10 nets are arranged in a line.

• Nets 1 through 5 were like super-strong fishing lines. They caught fish easily, even when the water was murky (low virus levels).
• Nets 6 through 10 were like flimsy, old string. They often broke or missed the fish entirely.

Because the second half of the puzzle (Nets 6–10) kept failing, the researchers could only reconstruct about 50% of the virus genome on average. They got a great picture of the first half, but the second half was often a blur or missing entirely.

Why Did This Happen?

The researchers asked: "Is it because the virus changed its shape (different genotypes)?"

• No. The problem happened with all types of HBV (Genotypes A, B, C, D, and E) equally.

"Is it because the virus was too weak (low viral load)?"

• Partly. When the virus load was very low (high "Ct value"), the weak nets failed even more often. But even when the virus was strong, the weak nets still struggled compared to the strong ones.

"Is it because the nets didn't fit the fish (mismatches)?"

• Not really. They checked the "nets" (primers) against the "fish" (virus DNA) and found they fit pretty well. The failure wasn't just about a bad fit; it seemed to be a fundamental design issue with those specific nets.

The Takeaway

1. It Works, But It's Lopsided: You can use this specific set of nets with the new Nanopore camera. It's fast and cheap.
2. The "First Half" is Great: If you just need to check for drug resistance or identify the virus type (which usually happens in the first half of the genome), this method is excellent.
3. The "Second Half" Needs Fixing: If you need the entire genome (for detailed research), this specific design is incomplete. The nets in the second half are just too weak.
4. Simplicity Wins: You don't need to do the complicated "two-bucket" method. Throwing them all in one bucket is just as good and saves time.

The Bottom Line

The researchers successfully adapted an old tool for a new, high-tech camera. However, they discovered that the tool has a "broken handle" on one side. While it's good enough for routine checks, scientists will need to design a new, sturdier set of nets to get a perfect, full-genome picture of the Hepatitis B virus every time.

Technical Summary

1. Problem Statement

• Clinical Need: Hepatitis B Virus (HBV) is a major global health burden. Effective management requires whole-genome sequencing (WGS) to monitor viral diversity, detect drug resistance, and identify recombination events, which partial sequencing (e.g., Sanger sequencing of specific genes) often misses.
• Technical Limitations:
  • Genomic Diversity: HBV has high genetic diversity (Genotypes A–E), making universal primer design difficult.
  • Platform Compatibility: A previously published tiling PCR scheme by Ringlander et al. (2022) was highly sensitive but designed specifically for Ion Torrent sequencing. It contained platform-specific adapter sequences and had not been validated for Oxford Nanopore Technologies (ONT) long-read sequencing.
  • Multiplexing Challenges: It was unknown if the Ringlander primers could function in a multiplexed format (all primers in one reaction) without primer competition or the formation of short, non-specific fragments, or if they required a two-pool strategy (non-overlapping amplicons) to function.

2. Methodology

The study evaluated the adaptation of the Ringlander et al. (2022) primer set for ONT sequencing using clinical samples.

• Sample Collection:
  • Source: 84 unique clinical serum/plasma samples from the Norwegian Institute of Public Health (NIPH).
  • Genotypes: Represented Genotypes A (30), B (19), C (16), D (37), and E (18).
  • Ct Range: Samples spanned a wide range of viral loads (Ct values).
  • DNA Extraction: Performed using Qiagen QIAamp or Abbott sp2000 systems.
• Primer Modification:
  • The original 10 primer pairs (producing ~400 bp overlapping amplicons) were modified by removing Ion Torrent-specific adapter sequences, retaining only the HBV-binding regions.
• Experimental Design (Pooling Strategies):
  • Strategy A (Single Pool): All 20 primers combined into one PCR reaction (20 µM total concentration).
  • Strategy B (Two Pools): Primers separated into two non-overlapping pools (Odd vs. Even amplicons) to avoid primer competition (10 µM total per pool).
  • Note: 36 samples were tested using both strategies to compare performance.
• Sequencing & Bioinformatics:
  • Platform: Oxford Nanopore GridION Mk1 with R10.4.1 flow cells.
  • Library Prep: ONT Rapid Barcoding Kit.
  • Pipeline: ARTIC Network pipeline (v1.8.5) adapted for Rapid kits.
  • Reference Handling: The circular HBV genome (NC_003977.2) was linearized and modified by duplicating the overlap region between amplicon 1 and 10 to prevent mapping fragmentation at the origin.
  • Analysis: Reads were classified using Kraken2, human reads removed via alignment to T2T-CHM13v2.0, and consensus sequences generated using Clair3 (minimum depth 20x).

3. Key Results

• Genome Coverage:
  • Overall Performance: The median genome coverage across all samples was ~50%. Coverage ranged from 0% (complete failure) to 99%.
  • Pooling Strategy Impact: There was no significant difference in genome recovery between the single-pool and two-pool strategies (Pearson correlation R = 0.68). Both strategies yielded similar results, suggesting primer competition was not the primary limiting factor.
• Amplicon-Specific Bias (The "Polarity" Effect):
  • High Performance: Amplicons 1–5 generally achieved high sequencing depths (median >100x for 2, 3, and 5).
  • Low Performance: Amplicons 6, 8, 9, and 10 consistently showed poor amplification (median depth <20x), often failing to meet the 20x consensus threshold. Amplicon 7 was intermediate.
  • Consequence: This uneven performance created a "coverage plateau" where the first half of the genome was well-covered, but the second half was frequently missing, limiting full-genome recovery.
• Impact of Viral Load (Ct Values):
  • Strong negative correlation between Ct value and genome coverage (R = -0.59).
  • Samples with low Ct (<20) achieved 50–80% coverage.
  • Samples with high Ct (>25) showed highly variable coverage, often dropping below 50%.
  • Resilience: Amplicons 3 and 5 remained robust even at Ct >30, whereas amplicons 9 and 10 failed across nearly all viral loads.
• Specificity:
  • High specificity was observed; >98% of reads were classified as HBV.
  • Non-HBV reads were primarily human DNA found in samples with failed PCRs (high Ct, low total reads), rather than off-target primer binding.
• Primer Mismatch Analysis:
  • Mismatch analysis across Genotypes A–E showed generally low mismatch counts (<1 on average).
  • While some poorly performing amplicons (4, 8, 9) had higher mismatch rates, the correlation was not consistent. Poorly performing amplicons 6 and 10 had low mismatch counts, suggesting that mismatches alone do not explain the amplification failure.

4. Key Contributions

1. Platform Adaptation: Successfully demonstrated that the Ringlander et al. (2022) primer set can be adapted for Oxford Nanopore sequencing by simply removing platform-specific adapters.
2. Multiplexing Validation: Confirmed that a single-pool multiplex PCR approach is feasible for this primer set, offering a streamlined workflow (one reaction per sample) without significant loss of performance compared to a two-pool strategy.
3. Identification of Design Flaws: Identified a reproducible "polarity" in amplification efficiency where the 3' end of the genome (amplicons 6–10) consistently underperforms, regardless of genotype or viral load.
4. Benchmarking: Provided a comparative analysis showing that while the Ringlander scheme works, newer designs (e.g., Lumley et al., 2025) with redundant primers per site offer superior uniformity.

5. Significance and Conclusion

• Feasibility: The study proves that tiling PCR for HBV WGS is feasible on Nanopore platforms using adapted primers, enabling rapid, cost-effective genotyping and resistance monitoring.
• Limitations: The current scheme is limited by uneven amplicon performance, resulting in incomplete genomes for many samples (median 50% coverage). This limits its utility for full-genome surveillance unless optimized.
• Clinical Utility: Despite incomplete genomes, the most clinically relevant regions (Polymerase and Surface genes, covered by Amplicons 1–3) are consistently recovered, allowing for standard genotyping and resistance testing even when full genomes are not obtained.
• Future Directions: The authors conclude that further optimization of the primer design is necessary. Specifically, redesigning primers for the underperforming regions (6–10) or adopting strategies with multiple primers per binding site (as seen in newer literature) is required to achieve consistent, full-genome recovery. The single-pool strategy is recommended for high-throughput workflows due to its efficiency, provided the amplicon bias is addressed.