Today's news brings a bold prediction from one of the leaders in genome sequencing, Oxford Nanopore Technologies. (Disclosure: David Deamer, a co-author of this book, is an adviser to ONT and played a key role in the company's breakthrough in using nanopores to sequence genomes.)
If you thought computing power developed rapidly, consider this from the conclusion of the article below: "the genomic revolution is moving three times faster than Moore’s Law at the least"
[posted by Wallace Kaufman]
Oxford Nanopore Expands Sequencing and CRISPR Applications
Oxford Nanopore (ONT) - a private company and potential competitor to Illumina, PacBio, and Qiagen - announced a number of groundbreaking developments in the field of sequencing technology this week. In addition to new pores, new chemistries, and updates on various devices, CTO Clive Brown detailed a CRISPR-based Cas9 enzyme application that the company plans to commercialize with mobile sequencers. Although it has overpromised and under-delivered for the past four or five years, ONT could become a disruptive innovator in the sequencing market if it is able to commercialize even half of the new products and services – importantly those based on CRSPR – it featured this week.
ARK believes that a CRSPR-based Cas9 application to enrich DNA in silico could be the next significant breakthrough in sequencing technology, causing another step function decline in time and costs. With a probe RNA molecule, ONT has disabled the Cas9 cutting activity and enriched DNA in silico in order to sequence specific regions of DNA: it calls this breakthrough “on-demand” sequencing. ONT also plans to commercialize pre-packaged probe kits for both protein detection and molecular diagnostic tests. That said, “on-demand” sequencing must make significant leaps in both specificity and sensitivity to become truly disruptive to existing platforms in the market during the next few years.
ONT’s pore technology is another candidate for disruptive innovation in this space, thanks to its base-calling accuracy. The R9.4 can now sequence 450 bases/second, an 80% improvement in speed without comprising accuracy, compared to previous generations. Using recurrent neural networks (RNNs) instead of the traditional Hidden Markov Model (HMM), ONT can train the sequencer to read chromatogram data more accurately into nucleobases (the sequence of nucleic acid in DNA), recognizing homopolymers. Oxford Nanopore’s signal analysis with neural networks is a fast moving new field with cutting edge algorithms. Combined with field-programmable gate arrays (FPGAs), RNNs could turbocharge sequencing efficiencies.
Brown revealed that one of its long term goals is to create a one million channel sensor which, at a 100% yield rate, could sequence a human genome in as little as three minutes. Today’s state of the art machines sequence the approximately ~3 billion base pairs of DNA in human genomes in one to three hours. Because the genomic revolution is moving three times faster than Moore’s Law at the least, ARK is keeping a close eye on this space!