This article has been updated with additional information from Biomodal.
NEW YORK – Illumina last week provided an update on its multiomics product and R&D pipeline at the American Society of Human Genetics annual meeting. During several company-sponsored workshops, the firm highlighted its "constellation" mapped read technology, Illumina Protein Prep proteomics assay, 5-base genome, PIPseq single-cell sample prep, and paired 500-base MiSeq reads.
In August, the company introduced technology it then called "comprehensive whole-genome sequencing," which moves library preparation onto the flow cell and offers long-range information from mapping short sequence reads onto larger DNA fragments.
Chief Technology Officer Steve Barnard explained that the approach, renamed "constellation mapped read technology," starts with long DNA fragments entering the flow cell and lying down across several wells. At those wells, the DNA is cut, tagged, and amplified to generate clusters for standard short-read sequencing using the company's XLEAP-SBS chemistry. From the distance between the wells, which form a pattern that resembles a constellation of stars in the sky when connected, the position of the reads on the original DNA molecule can be deduced.
Having the additional spatial information for the short reads helps increase sequence coverage because previously ambiguous reads that were discarded can now be mapped, Barnard said. It also helps with phasing variants and determining structural variants. The workflow only requires pipetting of three reagents onto a cartridge, he noted.
Illumina already has 10 early-access sites signed up for the technology, and early access will be expanded in the first half of 2025, according to Barnard. The company has also filed more than 45 patents on the technology.
The Broad Institute has been trying out the constellation protocol since this summer, working with scientists from Illumina's R&D team in the UK and the US. Niall Lennon, director of the Broad's genomics platform and CSO of Broad Clinical Labs, presented initial data and provided his first impressions during a company workshop. "We get offered access to many protocols all the time in our lab," he said, some more interesting than others. "This one was really exciting to try."
He and his team tried the technology on 40 test samples, including DNA extracted from cell lines, blood, and saliva. They were particularly impressed by the simplicity of the workflow, since library preparation happens entirely on the flow cell. According to Lennon, many other groups have promised to eliminate separate library prep in the past, but "this is the first time in 19 years I have seen someone actually come up with a method that has no library prep."
Lennon compared the long-range information from the constellation protocol to what could be gleaned from mate pair or jumping libraries of old, which required three days of sample preparation and 10 μg of input DNA. "You would get some jumping information, but it was really hard," he said, whereas the constellation protocol only requires 350 ng of DNA and no library prep.
For specific clinically important genes, read coverage increased significantly with constellation read mapping, he said, because reads could be mapped more accurately to regions that previously had low mappability. "You can actually start to recover variants that were incorrect in those low-mappability regions," he said.
His team just finished benchmarking the technology and found "our expected performance" for SNP calling across the genome in terms of precision, recall, sensitivity, and specificity, he said.
Large parts of the genome could also be phased, and structural variants could be called based on the spatial data. "Now you can actually confidently phase full genes because you have got this proximity information," he said, allowing researchers to look for compound heterozygous variants, for example.
While the percentage of phased genes dropped off with DNA from blood or saliva samples, compared to DNA from cell lines, "we still get pretty impressive numbers," Lennon noted.
Data analysis could also be further improved, he said, for example for calling structural variants. "This is pretty early technology, so a lot of the algorithmic methods to really take advantage of this data type haven't yet been developed," he said.
"It's very early, and we're just scratching the surface of what could be done with this data type," Lennon concluded, adding that the simplified sample prep and low input requirements make the technology especially interesting for "real-world applications."
According to Barnard, the constellation workflow can be run with up to 16 DNA samples per 24-hour run, with analysis provided by Illumina's Dragen software. Internally, Illumina has already processed more than 8,500 samples in more than 1,000 runs. In the future, the technology could enable sequencing from crude lysates, phased epigenetic sequencing, and nonhuman DNA sequencing, he added.
Proteomics assay
Illumina Protein Prep, the company's forthcoming proteomics assay developed in partnership with Standard BioTools, is powered by the so-called Slow Off-Rate Modified Aptamer (SOMAmer) technology, originally developed by SomaLogic.
Currently, the assay can analyze 6,000 unique human protein targets in parallel, a number that will increase to between 9,000 and 10,000 targets when it launches commercially in the first quarter of 2025, said Krishna Morampudi, Illumina's associate director of product management. The assay is also automatable using the Tecan Fluent 780 liquid handler, an OEM system that will be available through Illumina.
The library generated by the assay can be sequenced on the NovaSeq, followed by cloud-based analysis using Illumina Connected Analytics powered by Dragen, Morampudi said, adding that the data can be further analyzed using Partek's tertiary analysis software.
The protein prep workflow has a turnaround time of 2.5 days and offers femtomolar sensitivity as well as a 10-log dynamic range, according to Illumina.
Some researchers at ASHG appeared to be excited about the offering. "This is something I really would love to have," said Marilyn Li, director of cancer genomic diagnostics at Children's Hospital of Philadelphia. "I would like to be one of the earlier-access sites."
Currently, Li's group primarily runs immunohistochemistry (IHC) to analyze proteins in cancer samples, as well as some flow cytometry tests for hematological malignancies. However, she noted that IHC assays are not scalable and don’t meet her lab's needs. As such, she is excited to try out the Illumina Protein Prep, which is compatible with the three NovaSeq sequencers her lab operates.
Integrated methylation, variant analysis
Illumina's 5-base genome assay for DNA methylation analysis will go into early access at the beginning of next year, Joel Fellis, Illumina's VP of product management, told GenomeWeb at the conference. The assay can detect 5-methylcytosine and DNA variants in parallel, which he said is a differentiating factor from competing products such as Biomodal’s five-letter seq epigenetic analysis assays, which he said can only detect methylation signals. However, according to Biomodal, its Duet Multiomics Solution +modC and evoC assays can also detect methylation and genetic variants simultaneously in a single sequencing run.
Barnard said during the workshop that the 5-base genome workflow takes six hours and adds a single step to library preparation. Illumina's protein engineers "modified an enzyme to do something it doesn't want to do," he said, namely turning methylated C into T. The method is compatible with both PCR and PCR-free methods, he added, and comes with optimized data analysis software.
The 6-base genome, which will include another type of methylation, "is definitely something" the company is working on for the future, Fellis said. Down the road, Illumina also plans to combine its methylation assays with the constellation mapped read technology into a single assay, he added.
Single-cell sample prep
PIPseq V, the instrument-free single-cell sample prep product sold by Fluent BioSciences, which Illumina acquired this summer, is currently being integrated with Illumina's informatics tools and product portfolio, according to Fellis. The company will begin selling the kits, which are currently offered by the legacy Fluent team, under the Illumina name in the first quarter of 2025.
That first PIPseq release is "largely a repackaging" of the original assay, Fellis said, though the company plans to make further improvements. "We absolutely plan to build out a portfolio of single-cell solutions," he said. "We think there's a lot we can do to improve the overall performance of the technology, too."
During a company workshop, Shane Liddelow, a professor in the department of neuroscience and physiology at NYU Langone Health, called the technology "very scalable." He has been working with Fluent BioSciences and showed data from a single PIPseq V reaction that allowed him to sequence 200,000 single astrocytes from two mouse brains.
MiSeq 2x500 reads
Finally, Illumina is working on pushing read lengths for the recently launched MiSeq i100. According to Barnard, the company is running 2x500-base reads "routinely" now in its R&D lab, retaining read quality until it is "falling off a little bit" at the end. Generating such long reads has become possible through "tremendous improvements" in dyes, linkers, the protection group, and reagent stability, he said. "If anyone had told me 10 years ago we'd be doing 2x500 [base reads] with normal SBS chemistry, I would have said, 'probably not,'" Barnard said. However, he did not provide a timeline for launching long MiSeq reads as a product.