Identification of disease variants in a record timeframe using whole-genome nanopore sequencing
- Home
- Identification of disease variants in a record timeframe using whole-genome nanopore sequencing
Whole-genome sequencing (WGS) is the most comprehensive technique for identifying the variants underlying genetic diseases; however, the turnaround time required for such an approach is a barrier to its adoption for acutely ill patients. In a proof-of-principle study, a group of researchers, led by Stanford University School of Medicine, developed a novel workflow combining efficient sample preparation, ultra-rapid nanopore WGS, and accelerated variant calling and prioritisation1,2, demonstrating the ability to perform accurate genome-wide variant calling in clinical research samples in under 8 hours — the fastest turnaround time to date2.
The lengthy timelines associated with short-read-based WGS for pathogenic variant detection limit its potential application in settings where rapid data is required2. Goenka et al. presented an end-to-end workflow (Figure 1) for high-depth human nanopore WGS, requiring two hours of sequencing, in combination with real-time basecalling, sequence alignment, and accelerated small variant and structural variant (SV) calling and filtration. Optimising their approach on a human reference genome sample (HG002), the team demonstrated the application of this workflow to 12 clinical research samples1, achieving sample to identification of candidate causative variants in under eight hours2.
To achieve this, the team focused on efficient sample preparation, adapting the standard library preparation protocol to generate sufficient high-quality genomic DNA from a limited volume of blood, to allow distributed sequencing over all 48 flow cells of a PromethION 482. Per-sample cost could be reduced by reusing the flow cells for multiple samples, with a nuclease wash found to be sufficient between runs to avoid sample carryover. Within approximately two hours of sequencing, a high yield (200 Gb of sequence data) could be generated for downstream variant calling.
To reduce runtime, the team developed a cloud-based module (Google Cloud Platform) with massive parallelism and GPU acceleration, allowing high-accuracy basecalling and sequence alignment to be performed in real time whilst sequencing progressed. They also accelerated their variant calling pipelines. The researchers additionally noted that the utilisation of phase information was particularly valuable, with gains in variant calling performance clearly apparent for large and complex genes.
Figure 1. Overview of the ultra-rapid, end-to-end, whole-genome nanopore sequencing pipeline, from sample collection to candidate variant identification. Steps performed in parallel are vertically stacked. Using this pipeline, a candidate variant was found within eight hours for both samples studied. Figure adapted from Goenka et al. 20222.
Another challenge of traditional variant filtering and prioritisation approaches is the high number (~100) of candidates identified, which could potentially limit the future applicability of WGS in an ultra-rapid clinical setting; 'for application in the ultra-rapid setting, our goal was to readily surface clearly pathogenic, actionable variants in established disease genes'. The researchers therefore also focused on developing effective filtration and curation of candidate variants — the final stage in their ultra-rapid, end-to-end workflow (Figure 1).
Goenka et al. demonstrated the performance of their workflow in a real-world setting, applying it to two clinical research samples derived from a critically ill 57-year-old male with severe SARS-CoV-2 infection and comorbidities, and a 14-month-old female infant with history of dystonic/ opisthotonic posturing and developmental delay, for whom diagnostic testing proved inconclusive, suggesting possible genetic aetiology2. Demonstrating the power of their variant calling and filtration pipeline compared with the standard approach, for the neonate sample, filtration identified 31 small variants, and 21 structural variants, for manual review, compared with 147 variants via the standard system (Figure 2).
For both individuals, high-depth nanopore WGS was achieved in approximately two hours, variant calling and annotation were completed within seven hours from the start of sample preparation, and a candidate variant was obtained within eight hours.
Figure 2. Comparison of the standard and ultra-rapid variant filtration pipelines using a clinical research sample. With only 31 variants obtained via the ultra-rapid pipeline, compared to 147 variants found through the standard pipeline, the time required for manual review of candidate variants would be significantly reduced. Figure adapted from Goenka et al. 20222.
Gorzynski, J.E. et al. Ultra-rapid nanopore whole genome genetic diagnosis of dilated cardiomyopathy in an adolescent with cardiogenic shock.Circ. Genom. Precis. Med. 15(2):e003591 (2022). DOI: https://doi.org/10.1161/circgen.121.003591
Goenka, S.D. et al. Accelerated identification of disease-causing variants with ultra-rapid nanopore genome sequencing. Nat. Biotech. 40(7):1035–1041 (2022). DOI: https://doi.org/10.1038/s41587-022-01221-5