hen Illumina announced on Monday that it will begin shipping its two newest DNA sequencing machines as soon as March, president and CEO Francis deSouza said the technology “will enable the $100 genome.” That raised some questions:
$100 — really?
A $100 genome will cost $100 in the same way that the $1,000 genome costs $1,000. As in, it won’t, at least not soon. “The $1,000 genome” — which sequencer makers began promising about five years ago — “costs us $3,000,” said Richard Gibbs, founder of the Baylor College of Medicine Human Genome Sequencing Center and one of the leaders of the original Human Genome Project in the 1990s.
The headline price generally includes equipment and chemicals that get used up, labor, and depreciation of the machines, said Joel Fellis, Illumina’s director of product marketing for sequencing systems, “but it’s been really clear that our customers do the calculation differently.”
The truly important add-ons aren’t optional the way, say, extra leg room is with an airplane seat: A big chunk of the cost of meaningful human genome sequencing is for analysis, without which the raw sequence of the 3 billion A’s, T’s, C’s, and G’s is gibberish. In fact, when a manufacturer promises a $1,000 genome, even the raw readout — no analysis included — costs $1,200, said Michael Snyder of Stanford University, who sequenced and analyzed his own genome in 2012. Genome sequencing related to patient diagnosis and care — that is, in a clinical setting — is more, he said, with detailed interpretation costing up to $15,000.
That’s why Illumina isn’t making any formal claims about $100 genomes, said Fellis. “We’re not commenting on the exact math of that” $100 calculation, “but we’re very confident the architecture [of the new machines] will get customers close to that.”
OK, but when will it really cost $100 to get my genome sequenced and interpreted?
Sequencing and analyzing only the protein-coding regions of a human genome, called the exome, currently costs $1,000 to $5,000, Gibbs said. But if Illumina comes through with a “$100 genome,” even with all of the caveats above, “you can see a future where it costs a few hundred dollars” for exome sequencing and “you can see that it could be justified” for patient care, he said. “It gets us closer to where you can walk into your doctor’s office” and get your exome or even your full genome sequenced in a way that could provide actionable information.
Even if $100 is a few years off, “we’re talking about a cost low enough that any physician can order sequencing on any individual for any reason for less than the cost of an X-ray,” said Dr. Robert Green, an expert in genetic medicine at Brigham and Women’s Hospital. “It’s very exciting to hear the leading manufacturer of genome sequencers say they are pointing toward a $100 genome.”
Does the increased speed of the new machines matter in genome research?
The NovaSeq5000 and NovaSeq6000 systems (which cost $850,000 and $985,000, respectively) will be three times faster than the previous generation of Illumina sequencers, able to sequence 48 human genomes in the standard way (30X coverage, for you genomicists) in the time that current technology can do 16.
For studies looking for, say, rare genetic variants that affect people’s risk of developing common genetic disorders, it’s not unusual to need tens of thousands of genomes from people with and without the disorder, said Gibbs, whose center has ordered one of the 6000s for March delivery. “Just getting that genomic data is half the work,” he added. “So if [the new sequencers] can speed that up by a factor of two or three, that’s a big deal.”
Comparing the DNA of tumors to that of normal tissues also requires huge numbers of samples, so in those studies, too, speed can make a difference. But although “for thousands of samples speed can be a factor,” said Stanford’s Snyder, “budget is usually the biggest factor” in how many genomes a study can analyze.
Illumina is confident the extra speed matters to its customers. “Genomics is a very competitive space, and if you’re competing for grants or publications, speed is a huge advantage,” Fellis said. “And obviously there are applications where time is really important,” such as in identifying the genetic cause of a child’s mysterious disease. Tripling speed can also enable more labs to sequence tumor DNA to 300X coverage (meaning, roughly, that each genome is sequenced that many times, where higher “X” brings greater accuracy) rather than the current target of up to 90X. This can allow researchers to identify crucial cancer-driving mutations missed with lower coverage.
Will speed matter to patients?
Sequencing accounts for only a few days of the about two weeks it takes for a DNA sample to go through quality control, preparation, and actual sequencing, blogged bioinformaticist Mick Watson of Edinburgh Genomics. “So instead of waiting 14 days, you’re waiting 13 days. Who cares?”
With some sequencers turning around samples in a couple of days even now, the extra speed could in fact be significant, said Green, co-leader of the BabySeq Project to study the use of genome sequencing in newborns.
Who besides Baylor is buying the NovaSeq machines?
Illumina said it has orders for a total of 49 machines from, among others, the Broad Institute, Human Longevity Inc., Regeneron Pharmaceuticals, and the Chan Zuckerberg Biohub. That’s nearly double what Illumina says it can ship this quarter, but it expects to clear the order backlog later this year.