Oxford Nanopore announces multiple releases, for high-accuracy, content-rich, high-throughput whole-genome sequencing, and dynamic targeted sequencing

- announced at the 2020 American Society of Human Genetics (ASHG) Meeting.

Oxford, UK, 28th October 2020. Oxford Nanopore announces multiple releases that expand and improve its sequencing technology offering in human genomics, at ASHG 2020.

These releases are provided without upgrade fees, as with all Oxford Nanopore platforms.

They include: a range of data analysis releases, including new basecalling algorithms and downstream analysis tool upgrades.  These deliver increased performance across raw-read accuracy, single nucleotide polymorphism (SNP) calling and structural variant (SV) detection, alongside the improved ability to directly detect methylation in the same experiment.

Recent releases also support the use of PromethION, the ultra-high throughput nanopore sequencing device, in production-scale, cost-effective, whole human genome sequencing projects, where information-rich nanopore data enables broader genomic discovery.

Oxford Nanopore has also released two novel approaches to targeting specific genomic regions; adaptive sampling, and Cas9 Sequencing Kits for PCR-free target enrichment.

These announcements follow updates earlier in the year, including the release of the GridION Q — the first in a pipeline of ISO9001-qualified ‘Q line’ devices — and the CE-IVD marking and launch of its LamPORE COVID-19 test.

Oxford Nanopore will be hosting a satellite event today at ASHG, including presentations by Ariel Gershman of Johns Hopkins University, Tuuli Lappalainen of the New York Genome Centre, Shruti Iyer of Cold Spring Harbor Laboratory, and Rosemary Dokos from Oxford Nanopore. Talks will also be available online after the event at the ASHG and Oxford Nanopore websites.

Releases and improvements include:

Continuous performance iteration and accuracy

  • Single-molecule accuracy: A new analysis tool and protocol, as demonstrated by Karst et al, enables nanopore sequencing with unique molecular identifiers (UMIs) to produce 99.9% accuracy of an individual amplicon, with the ability to generate higher accuracy with higher coverage and reach lower frequency variants.  Karst et al showed 99.995% accurate single molecule consensus amplicons with the R10.3 nanopore[1].
  • Raw-read accuracy: A new basecalling algorithm, Bonito CRF, has been released to early users. This latest update to Bonito builds on previous work to deliver improved performance, and is trained with a larger, more diverse data set. This has improved modal raw-read accuracy to 98.3% on internal validation sets with a substantial fraction of raw reads above Q20. Further improvements are anticipated and Bonito CRF will also be used on R10.3 and UMI data.
  • High consensus accuracy: In early 2020, nanopore sequencing flow cells using the R10.3 nanopore, for improved homopolymer resolution, were released for the MinION and GridION devices.  R10.3 flow cells for use with the high-throughput PromethION device, or for rapid, low cost sequencing on the Flongle adapter, will be commercially available imminently.
  • Structural variation calling:  The release of “SV Pipeline 2.0” results in very high-performance SV calling, to 97.5% precision and 95.5% recall. This supports richer characterisation of genomic samples.
  • Variant-calling tools: The analysis tool Medaka has now been paired with an additional tool, DeepVariant, to identify SNPs with 99.92 recall and precision, a level that is comparable to traditional short read sequencing (SRS).
  • Variant-calling publications: New evaluations of single nucleotide polymorphism (SNP) calling from the scientific community, in particular by Moller et al, demonstrate consistent performance against the latest GIAB truth sets (datasets with a complete picture of known variants), achieving over 99% precision using a variety of different software pipelines. In challenging regions of the genome, this performance exceeds that of short sequencing reads by 2-4%, as well as “finding more than 50,000 variants, which are completely inaccessible to SRS... these regions include medically relevant genes like CFC1B”.
  • Datasets: To support a better understanding of the latest nanopore tools and sequencing data, Oxford Nanopore has released a human genome data set, HG002.  This demonstrates the Oxford Nanopore sequencing platform’s latest performance, and will be regularly updated as new novel algorithms, tools and chemistries are developed.
  • Oxford Nanopore will be presenting at the Nanopore Community Meeting, 1-3 December 2020, further updates on consolidation of all these recent innovations.

Information-rich data

  • Following continuous innovations that drive an improved ability to characterise methylated bases during the sequencing process, nanopore sequencing now gives strong correlation to whole genome bisulphite sequencing, even at low read depth of ~20X.  The analysis delivers a higher number of methylated positions called, better reproducibility than bisulphite sequencing, and is available as part of the standard nanopore sequencing process with no additional sample preparation or cost.
  • Oxford Nanopore devices continue to enable the direct sequencing of any length fragment of DNA/RNA, from short to ultra-long; the current record for longest fragment sequenced is 4.2 Mb.  Sequencing longer fragments provides rich data on structural variants, which account for the largest number of divergent base pairs across human genomes and contribute to multiple human diseases[2].

Targeted sequencing

  • The release of ‘Adaptive sampling’ on GridION enables real-time, intelligent target enrichment with no upfront sample preparation. A user programmes the MinKNOW software to target specific regions of interest.  Nanopore sequencing then enriches for these regions in real time, by rejecting reads from non-target fragments. This method retains read length and methylation information, for rich discovery.
  • A Cas9 sequencing kit was recently launched for targeted nanopore sequencing. This kit enables high depth sequencing of previously inaccessible target regions capturing read lengths of up to 100 kb, retaining methylation information.

High-throughput human genomes

  • As part of an ongoing programme of continuous innovation on PromethION flow cells, users can sequence a whole human genome per PromethION flow cell, to ~30X coverage, with a roadmap targeting two human genomes per flow cell in the coming weeks.  A single PromethION 48 device can run up to 48 flow cells at a time and 96 flow cells per week.
  • The release of LSK110, “Kit 10” this month enables higher sequencing yields through kit chemistry innovation.
  • This week, Oxford Nanopore has released, with sample prep automation specialists Hamilton, automation scripts to support high throughput sample preparation for the nanopore sequencing of 1,000 whole human genomes per week.

Find out more about nanopore technology for human genomics.