Sequencing DNA by converting it into an accordion molecule

The newest DNA sequencing technology from Swiss multinational Roche doesn’t measure DNA directly but in fact analyzes a different polymer altogether. The technology is not yet available for sale, but it is already generating a buzz. “Our approach to efficiently sequencing DNA was to not sequence DNA,” Mark Kokoris, head of the team behind Roche’s forthcoming instrument, said in a recent webinar introducing the technology. Instead, his team’s approach is to convert the DNA into a much-larger surrogate polymer that can be read with a nanopore.

Nanopore sequencing works by measuring differences in current as DNA passes through a pore embedded in a membrane. Nucleobases of different size and shape block that current to a different degree.

Companies like Oxford Nanopore Technologies and PacBio have used nanopores to enable long single-molecule sequencing reads. But nanopore sequencing of unmodified DNA generates complex, noisy signals that require a lot of processing, and sequencing by synthesis—offered by market leader Illumina along with a crop of smaller competitors—has dominated the market.

Kokoris came up with the idea of solving some of the noise issue of nanopore sequencing by converting DNA into larger surrogate polymers, and he cofounded sequencing company Stratos Genomics back in 2007 to do it. Roche acquired the company in 2020 and is preparing to launch its technology, which Roche refers to as sequencing by expansion.

The Stratos team designed a molecule that is copied from DNA but has larger monomers that cause more-dramatic differences in charge when they pass through a nanopore. Making the system work, says Kokoris in an interview with C&EN, “was just a matter of dissecting every part of the process.”

In a preprint article posted late in February, the researchers describe the chemistry behind their new polymer, which the company calls an xpandomer, as it stood in 2020 (bioRxiv 2025, DOI: 10.1101/2025.02.19.639056). Since then, Kokoris says in an interview with C&EN, they have continued to optimize and improve the system.

The monomers of the new system include a nucleobase and a triphosphate group, just like natural DNA monomers—but the resemblance stops there. The phosphate groups are linked to the base by a nitrogen that is not part of the backbone of natural DNA, and each monomer also includes a large hairpin-shaped reporter moiety fused to base and phosphate through click chemistry. “It’s essentially an expandable nucleotide triphosphate,” Kokoris says.

The researchers knew that the new polymer’s chemistry was far from the syntheses that DNA polymerases usually handle, so while developing the monomers, they also began engineering enzymes. Starting with a repair polymerase from a hot spring–loving extremophile, they used directed evolution to find an enzyme that could handle the altered backbone and unwieldy reporter side chain.

Kokoris says that after years of optimization, the enzyme Roche researchers are currently working with yields sequencing reads that are 99.35% faithful to the original DNA template sequence. By sequencing both strands of a DNA duplex, Kokoris says they can achieve higher accuracy.

In an email to C&EN, enzyme engineer Maximilian Fürst of the University of Groningen says he is impressed by the engineered polymerase—but its accuracy still falls short of that of natural polymerases that replicate DNA, which are typically at least 99.99% accurate. “There are plenty of engineered polymerases able to incorporate new chemistries, but there may be only very few or none that have been engineered all the way on that last mile to additionally exhibit native substrate-like fidelity,” he says.

After the polymerase synthesizes the new molecule, acid cleavage breaks the nitrogen-phosphorus bond in each monomer’s backbone. This allows the hairpin to unfold like an accordion being opened, which inspired the researchers to dub the polymers xpandomers. In its tightly pleated form, the xpandomer is about the same length as the DNA template strand; after unfurling, it’s about 50 times as long. As a result, only one monomer occupies the nanopore at a time as the unfolded molecule passes through.

Each hairpin includes an ester-rich side chain that the Roche team thinks sterically lodges in the nanopore, preventing the molecule from continuing through until they apply a subsequent voltage pulse. This lets the researchers control the speed of passage, sidestepping a common problem for nanopore sequencing, where DNA can slip through a nanopore in fits and starts that make it hard to read homopolymers.

Omics blogger and industry observer Keith Robison says that Roche’s new sequencer and preliminary data from the instrument were “the talk of the meeting” at the Advances in Genome Biology and Technology (AGBT) conference in Florida at the end of February. The company unveiled its first preview in a 2 h technical webinar on Feb. 20, just before AGBT began.

Through an early-access program, Roche has tested the instrument with geneticists at the Broad Institute of MIT and Harvard and the Hartwig Medical Foundation who are working on rare-disease diagnosis and cancer genetics. During AGBT, researchers from both institutes presented their findings.

Experts left the meeting with questions remaining about pricing, suitability for different applications, and whether the sequencing chemistry includes systemic errors. Despite those questions, Robison says, “everyone else in the industry is now coming up with war plans of how they’re going to compete against Roche.”