NEW YORK – Using nanopore long-read sequencing, researchers from Dubai Health Genomic Medicine Center have solved undiagnosed rare disease cases while uncovering a methylation signature that could potentially be harnessed for diagnosing spinal muscular atrophy (SMA).
Described in a Nature Communications paper published last Friday, the study demonstrates the utility of long-read nanopore sequencing in the clinical setting as well as its ability to uncover novel disease signatures using epigenetic signals.
"Our results demonstrate the potential of long-read sequencing as a single unified assay for routine clinical genetic testing and the discovery of novel rare disease variations," the authors wrote.
In an email, Ahmad Abou Tayoun, a researcher at Dubai Health and the corresponding author of the study, said his team has started validating nanopore sequencing for clinical use for several diseases, including thalassemias, spinal muscular atrophy (SMA), and muscle weakness.
Dubai Health also recently struck a partnership with Oxford Nanopore Technologies to promote next-generation sequencing for carrier screening in the region.
According to Abou Tayoun, when implemented as a clinical test, the nanopore assay, like other testing, will be covered by the government for local residents while expats will be covered through private insurance or have to pay out of pocket.
For their study, the Dubai Health team optimized a wet lab protocol for long-read whole-genome sequencing using Oxford Nanopore's PromethIon platform. The method targets a minimum of 30X coverage with an average N50 of 12 kb, the authors noted.
Following sequencing, the researchers also developed a "simplified funnel-down filtration strategy" that aims to informatically reduce the number of variants, boosting the identification of small and large pathogenic variants as well as abnormal epigenetic disease profiles.
More specifically, the researchers used a combination of pipelines for the detection, annotation, and selection of short variants — namely single nucleotide variants (SNVs) and indels — as well as genome-wide rearrangements, such as copy number variations (CNVs) and structural variations (SVs). They benchmarked this genome module with samples from 17 patients who had a confirmed genetic diagnosis and found it was able to detect all associated pathogenic variants in this group.
To analyze the epigenome, the researcher leveraged a tool called Epimarker, which focuses on epigenetic signatures specific to 36 Mendelian neurodevelopmental disorders (MNDD). The authors benchmarked this epigenome module with a dataset of 57 samples, including 17 patients with clinically confirmed abnormal methylation profiles and 40 controls. Of the 17 patient samples, nine had MNDD, according to the study.
Overall, the study showed that Epimarker classified all patients with 100 percent sensitivity, while none of the control samples were assigned to MNDD.
Moreover, the researchers observed a methylation profile across the SMN1 gene, which is associated with SMA, and its homologous pseudogene, SMN2. Based on this observation, they proposed a workflow "where methylation spanning SMN1 introns 6 and 8 can be used as a tag for SMA diagnosis and carrier status determination, which can then be confirmed upon long-read sequencing read deconvolution using SMN1 and SMN2 paralog-specific variants at this locus."
The researchers further validated their long-read workflow in a cohort of 51 undiagnosed patients with suspected rare disease who had inconclusive test results from short-read whole-exome sequencing.
Of these, 41 percent had undergone multiple genetic testing and 86 percent had received chromosomal microarray testing. Ninety percent of the patients were of Arab descent, and 44 percent of them were female.
With nanopore long-read sequencing, the researchers achieved an average of 49X coverage for these patient samples, with an N50 of 11.7 kb. Since all were previously tested by exome sequencing, the researchers focused their analysis on SNVs within exons that were previously missed and those in the 50 bp exon-intron boundary by assessing their splicing potential.
Overall, they concluded that their approach "substantially reduces the number of variants detected by whole-genome long-read sequencing while capturing a wide spectrum of genomic and epigenomic pathogenic variation." This led to diagnoses for five out 51 patients, or roughly 10 percent.
The authors also examined their previously discovered methylation profile for SMA across all undiagnosed patients and observed a methylation profile consistent with biallelic loss of SMN1 in one patient. Read deconvolution analysis as well as droplet digital PCR confirmed the SMA diagnosis in this patient, the study noted.
Despite the promises of their long-read clinical sequencing approach, the researchers also acknowledged the potential limitation of their funnel-down filtration analysis strategy. For instance, they noted that the method may have reduced sensitivity as a result of filtering out large SV events in the noncoding regions, which might be pathogenic.
While the study authors did not discuss the cost of their workflow, Abou Tayoun noted that the cost of nanopore sequencing "is not significantly higher than standard-of-care testing, while it actually generates more data and higher anticipated yield."