Article courtesy of Science
For DNA sequencing, this “is the year of the big shake-up,” says Michael Snyder, a systems biologist at Stanford University. Sequencing is crucial to fields from basic biology to virology to human evolution, and its importance keeps growing. Clinicians are clamoring to harness it for early detection of cancer and other diseases, and biologists are finding ever more ways to use genomics to study single cells. But for years, most sequencing has relied on machines from a single company, Illumina.
Last week, however, a young company called Ultima Genomics said at a meeting in Orlando, Florida, that with new twists on existing technologies, it could provide human genomes for $100 a pop, one-fifth the going rate. Several other companies also promised faster, cheaper sequencing at the same meeting, Advances in Genome Biology and Technology. This year, key patents protecting Illumina’s sequencing technology will expire, paving the way for more competition, including from a Chinese company, MGI, which last week announced it would begin to sell its machines in the United States this summer. “We may be on the brink of the next revolution in sequencing,” says Beth Shapiro, an evolutionary biologist at the University of California, Santa Cruz (UCSC).
Most sequencing companies, including Illumina, which has controlled 80% of the global market, depend on “sequencing by synthesis.” The DNA to be deciphered is separated into single strands, which are usually chopped into short pieces and mounted on a surface—often a tiny bead—in a container called a flow cell. Each single strand fragment serves as a template to guide the synthesis of a strand with complementary bases, supplied one at a time to channels of beads. Because each added base has been modified to glow, a camera can record where it attaches—and hence the identity of the corresponding base on the original strand. The steps are repeated until the new DNA strand is complete.
Ultima streamlined the process by spraying the DNA-laden beads by the billions onto round silicon wafers the size of dessert plates. Nozzles above each wafer gently squirt out bases and other reagents, which spread thinly and evenly across the wafer as it rotates, reducing the amount of these expensive materials needed. Instead of moving back and forth across under the camera, the disk moves in a spiral, akin to how a compact disk is played, which speeds up imaging. It’s “clever engineering [that] avoids a lot of complex plumbing,” says Mark Akeson, a molecular biologist at UCSC. A neural network program rapidly turns imaging data into a sequence.
The sequencing chemistry is different as well. Only a few bases carry fluorescent tags, reducing costs. Moreover, the bases lack the usual stop signal, which ensures no extra bases latch on. Without these “terminators,” the growing chain can sometimes add multiple bases at once, speeding the process. “Many of these innovations are used elsewhere, but they seem to have come together very nicely here,” says Jay Shendure, a geneticist and technology developer at the University of Washington (UW), Seattle.
Ultima CEO Gilad Almogy and his colleagues demonstrated the technology’s potential in four preprints posted in late May on bioRxiv. In one, they and colleagues at the Broad Institute of MIT and Harvard used their machine to sequence more than 224 already-sequenced human genomes and found their results on par with previous work. The three other studies showed the technology can evaluate a single cell’s repertoire of expressed genes, the effects of mutations, and epigenetics—chemical modifications of DNA that affect gene activity.
Until now, cost has limited such single-cell studies, causing a bottleneck in research. But Snyder found Ultima’s low-cost approach enabled him to sequence multiple colon cancer cells to document how one DNA modification, methylation, changes as colon cancer develops.
In another preprint, Joshua Levin and his Broad Institute colleagues tested the ability of the Ultima technology to identify active genes in single blood cells as indicated by the genes’ RNA transcripts. The team found Ultima’s machine identified those genes about as well as Illumina’s did. And, he adds, “It’s a game changer due to the lower cost.”
Florence Chardon, a UW genomics graduate student who modifies DNA with the genome editor CRISPR, is excited by that prospect. “The less expensive [sequencing] gets, the more accessible this kind of research is to more labs and more people,” she says.
But Lior Pachter, a computational biologist at the California Institute of Technology, has reservations about the new technology. He and graduate student A. Sina Booeshaghi looked at one of the most active genes in blood cells from Levin’s team, a possible cancer biomarker also known for producing a protein athletes sometimes inject to illegally enhance their performance. The Ultima technology sometimes missed the active gene, Pachter says. The “error rate was very high, and the performance was very poor.”
The gene has a stretch in which the same base is repeated eight times, and Ultima admits long repeats can undermine the accuracy of its reads. Looking elsewhere in the Ultima sequence, Pachter found errors when one base was repeated just three times. He notes that a human genome contains at least 1.4 million of these so-called homopolymers. Still, he says, “For some applications, you don’t need perfect sequences.”
Pachter and others also take issue with the touted $100 cost. That figure only covers reagents, not the labor, pre- and postsequencing steps, and initial outlay for the machine, the price of which has not been released. Even if the $100 figure is real, it may not be unique: Other companies are also promising $100 per human genome.
One is MGI, a subsidiary of Chinese sequencing giant BGI. MGI’s technology is similar to Illumina’s, but it increases accuracy by adding all four bases at once as it sequences DNA. To track which bases are incorporated, it uses antibodies, which are brighter and less expensive than fluorescent dyes. Illumina, too, is promising lower costs, and at the meeting it introduced new chemistries to increase accuracy and flexibility.
For this bargain rate to be realized, Ultima and MGI both require filling their sequencers to capacity with hundreds of genomes. But high-throughput sequencing “is not always good for clinical practice even if it is good economics,” says Greg Elgar, a genome biologist at Genomics England, because sometimes a physician needs just one or a few people’s genomes analyzed. Other companies with new flow cells and chemistries can economically sequence small numbers of genomes. At last week’s meeting, Element Biosciences CEO Molly He reported the company is now shipping benchtop sequencers that can sequence three human genomes at a time, at a cost of $560 each. Another company, Singular Genomics, also promises benchtop technology that doesn’t require high throughput for cost savings.
These machines, like those from Illumina, MGI, and Ultima, all decipher short fragments of DNA. But for the past 7 years, two companies, Pacific Biosciences and Oxford Nanopore Technologies, have worked on sequencing “long reads,” thousands of bases long, which leave fewer partial sequences to piece together into a full genome. The technologies “can sequence the native DNA molecule, in all its glory,” Elgar says. They have struggled with low accuracy and high cost, but he says they are on their way to becoming practical tools.
Don’t count the sequencing giant Illumina out just yet. Its scientists “probably have kept a couple of cards in their back pocket” to keep their position in the market strong, says Albert Vilella, a bioinformatician and genomics consultant in Cambridge, England. Nonetheless, Illumina faces unprecedented competition, he adds. “It’s time to look at the [DNA sequencing] landscape with fresh eyes.”