This story has been updated to correct that Illumina launched the TruSeq synthetic long read kit in 2014 based on its 2013 acquisition of Moleculo.
LOS ANGELES – Illumina revealed more about its bid to break off a chunk of the long-read sequencing market on Thursday but continues to withhold details of how its Complete Long-Read technology works.
Speaking to a packed room at a sponsored workshop at the American Society of Human Genetics annual meeting on Thursday, Andrew Shaver, Illumina's product manager, spoke about the firm's launch plans for the CLR technology after briefing the crowd about the recently launched NovaSeq X series.
"We don’t think this is going to replace short-read [sequencing] at all," Shaver told the audience. "Complete long reads are really just to accelerate access to that last bit of the genome that's challenging."
At present, the technology generates reads with an N50 — a metric used in genome assembly to describe the minimum size of contigs covering half the genome — of about 6 kb to 7 kb, while the N50 for phase blocks is about 200 kb.
DNA input requirements are flexible, ranging from 5 ng up to 1.2 μg, with a target input of 50 ng. Additionally, he said the turnaround time for the library prep process, which is also automation friendly, is between six and eight hours, with only a thermocycler required. When it comes to data analysis, complete long-read data can be processed by the built-in algorithm on Illumina's Dragen platform.
After briefly showcasing some application examples of the technology — resolving challenging genes, such as STRC, CATSPER2, CHRNA7, and NCF1, as well as helping to detect insertions and deletions — Shaver named several early customers of the Complete Long-Read technology, including researchers from Washington University in St. Louis, Stanford University, and the n-Lorem Foundation.
Library prep is a "critical part of the technology right now," Shaver said, but beyond that, he did not offer much clarity on what's going on inside the reaction tube.
"I want to actually understand the mechanism a little bit more," said George Charames, director of Mount Sinai Hospital's Advanced Molecular Diagnostics Laboratory, who attended the workshop. "This is the core question; that's what I want to know."
Illumina announced the Complete Long-Read technology in January, referring to it by its internal codename at that time, "Infinity." Throughout the year, Illumina officials have touted the new product without offering many technical details. Also, at last month’s Illumina Genomics Forum, CEO Francis deSouza revealed the firm's plans for two initial kit sizes and said a launch is planned for the first quarter of 2023, emphasizing that Illumina considers CLR to be a single-molecule technology.
"It’s not a synthetic long-read technology," deSouza said last month. "It is a complete and accurate representation of the genome at the single-molecule level. You get contiguous long reads to close key gaps."
While details of the CLR technology are still unknown, the last decade has seen some failed attempts of technologies trying to synthesize or stitch together short reads into longer ones. Illumina, for instance, launched the TruSeq synthetic long read kit in 2014 based on its 2013 acquisition of Moleculo. The company said the synthetic long read kit is no longer an active product offered to customers. Similarly,10x Genomics phased out its Linked Reads assay in 2020.
Some observers believe that CLR is based on intellectual property from Longas Technologies, an Australian startup that has been silent since announcing its mutagenesis-based "virtual long read" method in 2019.
At ASHG, Shaver presented results from a collaboration with Stanford professor Euan Ashley, where Illumina processed 14 samples of undiagnosed or complex diseases with CLR that were sequenced on the NovaSeq 6000.
The results showed an average N50 just over 5 kb, Shaver said, while some reads were longer than 20 kb. In addition to fully phasing the TNXB and CHD6 genes, CLR verified all variants that were previously identified using existing long-read technologies in these samples.
Shaver also shared that the n-Lorem Foundation, a California-based organization dedicated to studying ultra-rare diseases, used CLR to help design antisense oligonucleotides (ASOs) for personalized medicine.
Charames' Mount Sinai lab uses the NovaSeq 6000 for clinical sequencing, and he often has to deal with pseudogenes, large structural variants, intronic rearrangements, and other challenging bits of the genome that cannot be easily captured by short-read sequencing. CLR is "potentially game-changing," he said, if it works as promised in customers' hands, is cost-effective, and can be covered by insurance.
However, he said he remains unsure whether the technology produces synthetic long reads or native long reads. Theoretically speaking, synthetic long reads carry the potential for read dropout and biased results compared to native long reads, he added.
Answering follow-up questions from attendees after the presentation, Shaver dropped some hints about the inner workings of the technology. 'What we are doing is, when we have a long read, we have a number of short reads with marks, and we are merging them together using those marks," he explained.
Regarding the technology's performance on repeat regions, Shaver acknowledged that repeats tend to be one of the "most challenging" categories for the technology to tackle. "We are still in development, we are looking at those, but we are not going to get 100 percent," he added, referring to CGG repeats.
Shaver did not discuss the technology's cost during his presentation but acknowledged that a complete long-read whole genome will not have the same cost as a genome generated with short reads, mainly due to the over-sequencing requirement. According to him, the company has not even priced the technology internally yet.
When asked if complete long reads are synthetic, Shaver said the technology leverages Illumina's existing platform and core consumables. "It's a novel technology and still in development; we'll have more details in the near future," he responded.
Andrew Han contributed reporting to this article.