Rapid and Accurate Detection of SARS Coronavirus 2 by Nanopore Amplicon Sequencing

Xiaoxiao Li, Chao Li, Pengcheng Du, Shaoyun Li, Le Yu, Zhiqiang Zhao, Tingting Liu, Cong-kai Zhang, Sen-chao Zhang, Zhuang Yu, Chao-ran Dong, Qinggang GeFrontiers in Microbiology 2022 · DOI

TL;DR

Researchers have demonstrated that Nanopore Amplicon Sequencing (NAS) significantly outperforms the "gold standard" RT-qPCR in detecting SARS-CoV-2, particularly in samples with low viral loads or those stored long-term. By tiling the entire viral genome, NAS achieves 96.5% sensitivity and 100% specificity, providing a robust diagnostic tool that also offers critical genomic surveillance data in under 15 hours.

What Problem Does This Paper Address?

While Reverse Transcription quantitative Polymerase Chain Reaction (RT-qPCR) has been the primary tool for COVID-19 diagnosis, it isn't perfect. It is prone to false negatives caused by viral mutations in primer-binding regions, low viral titers, or degradation of RNA in stored samples.

Standard Next-Generation Sequencing (NGS) can solve these issues by looking at the whole genome, but it is often too slow, expensive, and computationally intensive for rapid clinical use. The authors sought to evaluate whether Nanopore Amplicon Sequencing (NAS)—a method that uses long-read sequencing on portable devices—could bridge the gap by providing high-accuracy detection and viral load estimation faster than traditional NGS.

Methods

The study employed a parallel detection strategy on 86 clinical samples (pharyngeal swabs, sputum, and fecal matter) and 27 negative controls.

  • The NAS Workflow: The team used the ARTIC protocol, which involves multiplex PCR to create overlapping "amplicons" (DNA fragments) covering nearly the entire SARS-CoV-2 genome. These were sequenced on the Oxford Nanopore Technologies (ONT) MinION platform.
  • Bioinformatics: Raw reads were filtered for quality (NanoFilt) and mapped to the Wuhan-Hu-1 reference genome (Minimap2). A "positive" NAS result was defined based on reference genome coverage (the percentage of the viral genome successfully sequenced) rather than just the number of reads.
  • Comparison: These results were benchmarked against routine RT-qPCR targeting the ORF1ab and N genes.

Datasets

The study used a specialized clinical dataset collected between February and May 2020 at a hospital in Wuhan:

  • 86 Positive Samples: 54 pharyngeal swabs, 31 sputum samples, and 1 fecal sample from confirmed COVID-19 patients.
  • 27 Negative Controls: Samples from patients who tested negative for COVID-19.
  • Reference Genome: SARS-CoV-2 isolate Wuhan-Hu-1 (Accession no. NC_045512.2).

Notably, many of these samples had been stored for six months before the NAS testing, providing a rigorous test of the method's sensitivity to degraded RNA.

Key Insights and Ideas

  • Coverage Over Count: The most important "aha moment" was that reference genome coverage is a far more reliable diagnostic indicator than the raw number of viral reads. Using a coverage cutoff of ~9%, the researchers could perfectly distinguish between true positives and background noise/contamination.
  • Rescuing "False Negatives": NAS successfully detected the virus in 17 samples that RT-qPCR missed. This was largely attributed to NAS's ability to detect fragments across the entire genome, making it more resilient to RNA degradation and primer-site mutations.
  • Sputum as a Severity Marker: The study found that viral loads in sputum (estimated via NAS coverage) were more closely related to disease severity and duration than those in pharyngeal swabs.
  • Speed vs. Depth: While RT-qPCR is faster (~2.5h), the NAS workflow (under 15h) is significantly faster than traditional Illumina-based NGS while providing the same level of genomic detail.

Key Results

  • Sensitivity: NAS achieved 96.5%, compared to just 80.2% for RT-qPCR in this specific cohort.
  • Specificity: NAS hit a perfect 100%, while RT-qPCR was 96.3% (due to one false positive in the control group).
  • Correlation: There was a strong correlation ($R^2 = 0.382$) between the NAS genome coverage and the RT-qPCR cycle threshold (Ct) values, proving that sequencing depth can serve as a proxy for viral load.
  • Turnaround Time: The entire NAS process, from sample to result, took less than 15 hours, with the sequencing portion taking only 8 hours.

Limitations and Future Work

The authors acknowledge several hurdles for widespread adoption:

  • Cost and Complexity: NAS currently carries higher per-sample costs and requires more specialized laboratory expertise than qPCR.
  • Sample Size: The study was a single-center cohort with a relatively small number of samples.
  • RNA Degradation: While NAS performed well on 6-month-old samples, the authors noted that the RNA was likely degraded, which prevented them from obtaining complete (100%) genome sequences for all samples.

Future Work: The researchers suggest NAS is best suited as a supplementary method—particularly for investigating suspected false negatives, tracking new variants (like Omicron), and performing genomic surveillance in real-time during outbreaks.

Summary updated

Abstract

Our results showed the high sensitivity and specificity of the NAS method for SARS-CoV-2 detection compared with RT-qPCR. The sequencing results were also used as an indicator of the viral load to display the viral dynamics during infection. This study proved the wide application prospect of nanopore sequencing detection for SARS-CoV-2 and may more knowledge about the clinical characteristics of COVID-19.
BibTeX
@inproceedings{li2022rapid,
  author = {Xiaoxiao Li and Chao Li and Pengcheng Du and Shaoyun Li and Le Yu and Zhiqiang Zhao and Tingting Liu and Cong-kai Zhang and Sen-chao Zhang and Zhuang Yu and Chao-ran Dong and Qinggang Ge},
  title = {{Rapid and Accurate Detection of SARS Coronavirus 2 by Nanopore Amplicon Sequencing}},
  booktitle = {Frontiers in Microbiology},
  year = {2022},
  doi = {10.3389/fmicb.2022.735363},
}
Figure