A Worthy Sequel: PacBio's New Sequencing System

PacBio just released the world's best whole genome sequencer. Will it be enough to keep the company afloat?

By Aaron Krol

October 1, 2015 | This Wednesday, in a surprise announcement, Pacific Biosciences of Menlo Park, Calif., confirmed rumors that it has been working on a smaller, more price-effective version of its RS II gene sequencer. But rather than push out a scaled-down benchtop instrument for simple use cases, as many had anticipated, the company unveiled a machine that improves on the RS II in every particular: less than half the cost, a third the size, and most importantly, almost seven times as powerful.

The new instrument is called the Sequel System, and PacBio expects to ship its first ten before the end of the year.

2015 has been a big year for new sequencing instruments, with every company that actually has a functional gene sequencer bringing something new to market. Illumina added the HiSeq 3000 and 4000, filling some gaps in its catalogue between the gargantuan HiSeq X Ten and the well-loved HiSeq 2500. BGI, which had previously offered its Complete Genomics technology only as a service, introduced the Revolocity system. Thermo Fisher refocused on what it does best, producing an Ion S5 series that will appeal to labs running high volumes of targeted gene panels. And while the idiosyncratic Oxford Nanopore Technologies has shrugged off the idea of traditional market launches, it did open an access program for its first high-throughput instrument, the PromethION.

Every one of these instruments (with the possible exception of the Revolocity) can reasonably claim to be the best at something. PacBio’s new Sequel System can’t compete on data volume with the HiSeq line or Complete Genomics, and won’t come close to the Ion S5 or Illumina MiSeq on price. It doesn’t have the wow factor of the Oxford Nanopore MinION’s extreme miniaturization.

But for the gold standard of high-quality, whole human genomes, PacBio has just released the best sequencer money can buy.

The reason I feel comfortable making that claim without seeing any data from users — indeed, without any specific data from PacBio — is that the company already had the best sequencer for whole genomes, in the RS II. Based on the same SMRT (single molecule, real time) technology, the Sequel will have all the same advantages: a random error profile that provides very high consensus accuracy; built-in DNA methylation data; and extremely long reads that let users assemble entire genomes from scratch, revealing the large structural changes to the genome that get washed out on other sequencing platforms.

PacBio’s problem has always been cost. The RS II sold for $750,000, in the same range as the much higher-volume HiSeq series. A deeply sequenced whole human genome on the RS II would run to tens of thousands more in consumables. The Sequel, however, will be priced at $350,000, a little closer to a mid-range instrument like the Illumina NextSeq; and while its consumables will actually cost more per run than the RS II, the much greater volume of data generated more than makes up the difference. In a call to investors this morning, PacBio CEO Mike Hunkapiller said that the Sequel should deliver a 10x human genome in one day, at a consumables cost of $3,000. A really high-quality 30 or 50x genome would be proportionally more.

That’s more than five times the speed of the RS II, for less than half the cost. PacBio is still far from the cheapest option for labs looking to sequence genetic material, but it’s no longer a wild outlier, a fact that should let its technology’s inherent advantages shine.

Price Premium

PacBio, however, is still not in an enviable position. The company has always struggled to turn a profit. For investors, a key point about the Sequel will be that it can be manufactured for a quarter the cost of the RS II, giving PacBio its first really strong profit margin. But even the Sequel won’t escape the trap of being a niche product, for users who want to study structural variation, run complete de novo assemblies, or look into a few developing fields like epigenetics and alternative splicing of RNA. At best, the Sequel changes the calculation for lab heads who might have liked to lean in these directions, but have balked at the cost.

It feels unfair that PacBio has been left at the margins of the booming genomics industry. Like Ion Torrent (now owned by Thermo Fisher), PacBio overpromised in its early days and eventually had to accept that it was not catching up with the dominant Illumina on raw sequencing power, forcing a pivot to narrower indications where it could compete. But unlike Thermo Fisher, for PacBio that process involved calling out real and fundamental problems with the direction of genomic science. The deluge of fast, cheap, short-read technology had blinded the field to whole categories of genetic variation. Even if only out of sheer self-interest, the PacBio team has done as much as anyone to turn that around.

Long-read sequencing on the RS II has moved quickly over the past year, from a first whole human genome assembly in November 2014, to a pipeline today that can produce them on demand. Along the way, PacBio has fostered a body of software tools adapted to long reads and de novo assembly, and helped create a market for other companies like BioNano, RainDance, and 10X Genomics to build instruments that illuminate the structural complexities of the genome. This renewed interest in seeing the entire genome clearly is good for science as a whole.

The Sequel might finally let PacBio reap the profits from this field it’s helped to sow. The new instrument, backed by the same sample preparation and software pipelines as the RS II, will be immediately supported for all the same applications as its predecessor, giving this project maximum reach for any scientists who could be won over by PacBio’s new, much more attractive price point.

The PacBio Sequel System, still a full-size instrument but not longer the behemoth RS II. Image credit: Pacific Biosciences

Nonetheless, there are still serious headwinds. Running a Sequel at scale remains a costly option, a sort of luxury genomics. What’s more, PacBio has genuine competition these days. Clever workarounds like 10X Genomics’ GemCode Platform make it possible to get something very like long reads out of an Illumina machine.

And that’s nothing next to what’s coming from Oxford Nanopore. No one knows if the company’s radically new technology will overcome all its accuracy issues, but experiments with the handheld MinION sequencer have produced some seriously cool results (often with help from software tools originally written for PacBio reads). The PromethION, essentially a battery of 48 MinIONs rigged together, promises to ramp the MinION’s ultra-low-cost long-read data up to the level of human genome sequencing — and if it’s successful, the technology has plenty of room to keep scaling up.

In short, the Sequel breaks new ground on the accessibility of true whole human genome sequencing, but it may not be on top for long.

A Toe in the Clinic

Still, PacBio has a few tricks up its sleeve. Unlike the RS II, Hunkapiller told investors today, the Sequel is built with a flexible architecture that will let the company keep designing new SMRT cells with more waveguides, the rate-limiting structures for sequencing. Not only does the Sequel jump to one million waveguides, compared to the RS II’s 150,000, but users can now expect to see that number rise. In the fast-advancing world of genomics, that ability to keep scaling is crucial. 

Perhaps even more importantly, PacBio is on track to become a force in diagnostics, a huge market that has barely been touched by sequencing companies. The Sequel is the fruit of a diagnostics-focused partnership with Roche, whose dogged determination to be a part of the sequencing revolution is unmatched anywhere else in pharma.

Roche’s efforts in genomics have been a bit hapless to date — the company has notably fumbled an attempt to buy Illumina, shuttered its subsidiary 454 Life Sciences, and abandoned a nanopore sequencing project — but with the Sequel, Roche has helped to add real value to its partner’s technology. Now, Dan Zabrowski, head of Roche’s sequencing unit, says that the Sequel will become the basis for a series of clinical launches beginning in late 2016, including an instrument that will aim for FDA clearance.

Neither partner is sharing what indications that program will focus on, but Huntington’s disease, fragile X syndrome, and a few other rare genetic conditions caused by structural variants are a good bet. And then there’s cancer, where there’s a huge need for a clinical instrument that can illuminate the large, compound mutations that in many cases drive tumors out of control.

These are all areas that, at least in the research lab, PacBio instruments have helped shed a great deal of light on. Whether the company will get a chance to bask in that light depends in large part on whether Roche can secure the first mover advantage in the clinic — and who’s coming up behind.