The buzz leading up to this week’s Consumer Genetics Show in Boston suggested that a major announcement would be made by the CEO of genomics technology provider Illumina, Jay Flatley. Illumina provides the most popular second-generation sequencing instrument currently on the market, the Genome Analyzer II, and has been making noises about moving into the personal genomics industry since at least the beginning of the year, so the announcement itself was not exactly a shock: Illumina is launching a personal genome sequencing service.
The launch comes with a new website, the appealingly titled everygenome.com
. The website offers clear explanations of the processes of genotyping and DNA sequencing targeted towards consumers, doctors and scientists, but little in the way of details on the service. Fortunately the coverage of the conference (and Flatley’s announcement) on Twitter
was extraordinarily intense, and many of the details in this post come from tweets by Dan Vorhaus
, Jason Bobe
, Matthew Herper
, the Genetic Alliance
and several others. There are also longer stories on the launch up at GenomeWeb
and Bio-IT World
, and Misha Angrist has a nicely concise summary
of the important details from a consumer’s point of view.
So what are Illumina actually offering here?
What you get for your money
The service will consist of complete genome sequencing to a depth of 30X (meaning that each base in the genome is sequenced an average of 30 times), using Illumina’s paired-end short-read technology. Coverage of 30X with short-read sequencing is reasonable; it will miss around 10% of the genome that is too repetitive to span with short reads, but in the remainder it will call most single-base variants with >99% accuracy, and will do a fairly good job of picking up small insertions and deletions and larger rearrangements (so-called structural variants).
Illumina’s technology is extremely well-established, and serves as the backbone for most large-scale genome sequencing projects currently underway (including the majority of the samples sequenced as part of the 1000 Genomes Project
); that gives it an edge over the more experimental technology employed by competing sequence provider Complete Genomics
Most importantly, the widespread use of Illumina technology in research means that there are robust solutions for many of the informatics challenges associated with stitching a genome together and deciding which genetic variants are real. Customers can thus feel confident that a genome generated using the Illumina platform will be of a high standard – although, as Illumina notes in its advice to clinicians
, any potentially medically relevant findings should be followed up with a formal, clinically validated genetic test.
The big surprise for me in this announcement is the price – or rather, the fact that the price isn’t actually all that surprising. At $48,000, the cost represents a halving of the price of the current best commercial offering on the market ($99,000 from Knome), but it’s still way out of the price range of most of us – and substantially above the $10-20,000 figure that Flatley discussed back in February as the price its first customers would pay. This is incremental inching forward rather than a great leap towards the democratisation of genome sequencing.
The goal here obviously isn’t to provide a universally available product, but to attract wealthy early adopters to subsidise the development of the infrastructure required for more widespread personal genomics, and to start breaking down the social barriers against adoption. In the GenomeWeb piece
Flatley argues that his goal is to “begin to open the discourse about how we are going to do this in the future, and use it medically within the appropriate regulatory and ethical infrastructures”. This is a worthy ambition, but it’s not quite what I was hoping for in this launch.
An opportunity yet to be seized
Sequencing a whole genome with short-read technology is difficult, but the real challenge is the next bit: translating the assembled sequence into useful information. Illumina will not be doing the interpretation of genome sequences themselves, but will instead draw on the expertise of relatively well-established personal genomics companies
23andMe, deCODEme, Navigenics and Knome. Customers will apparently have the option of choosing which of these providers they go through for their analysis, who can then add their own additional fees for this crucial stage.
I can see why farming out interpretation is an attractive option, given Illumina’s major strength in sequencing technology and the enormous investment in databases and people required to build up a complete interpretation service, but in focusing on the sequence generation side Illumina (like Complete Genomics before them) is restricting itself to the least attractive segment of the personal genomics market
. Sequence generation is a competitive space dominated by a single bottom line: whoever can generate an accurate genome sequence more quickly and cheaply than its competitors wins, and there is no room for complacency; there will always
be new technologies popping up over the horizon. Illumina has carefully invested in third-generation technologies
to keep its edge, but there’s no guarantee that this will last.
Interpretation, on the other hand, is a niche with nuance: here there is more than one axis of competition, and potentially a much slower decline in profit margins. At current rates of progress it is only a matter of years before generating a genome sequence costs in the tens of dollars; since most people will only need their genome sequenced once* very cheap sequencing will rapidly result in a depletion of the sequence generation market. Interpretation, on the other hand, is the gift that keeps on giving: as new research uncovers more details about the functional portions of the genome and their interactions with environmental risk factors customers will always need their reports updated and re-analysed in increasingly more sophisticated ways.
Right now Illumina has both the reputation and the technological edge it needs to dominate the sequencing market at this crucial period in the development of personal genomics. The company now has a choice: it can farm out the interpretation service, focus on sequencing, and then rely on being able to always stay on top of a chaotic and ruthless arms race; or it can seize the opportunity to invest heavily in its own interpretation infrastructure, build a reputation as a complete genomics provider and then work to maintain that status even if it loses its edge in sequencing to an upstart competitor.
Obviously this isn’t a trivial decision for a business to make, and Illumina has plenty of other irons in the fire (in addition to dominating the sequencing market, Illumina also does a thriving business in providing the genotyping chips used for genome-wide association studies). However, I’d personally love to see the company shoulder its way into the interpretation space, even if only to keep the competition churning in this important area; perhaps the iPhone app previewed at Flatley’s talk (discussed below) is a small step in this direction.
* Added in edit: In the comments, Carl Shulman correctly notes that repeated sequencing will be useful for reasons other than generating a base genome sequence; of these, I’d suggest that the only application likely to become a widespread clinical tool in the next few years is cancer sequencing; things like transcriptomics (for looking at gene expression) and methylomics (for looking at patterns of DNA modification) have a much longer time horizon due to the complexity of biomarker validation. There certainly is a bright future for sequence-based diagnostics in general, but note that in all cases the real value comes from interpretation rather from sequencing per se, so I think a similar argument applies across the board for the benefits of entering the interpretation market.
Speaking of competition in the interpretation arena: it will be fascinating to see how the four current interpretation providers fare in attracting Illumina’s sequencing customers. Knome has an obvious edge in many ways, being the only member of the four with actual experience in dealing with complete genome sequence data, but it’s unclear how much the company will charge as its interpretation fee.
23andMe, on the other hand, has no direct experience in explaining whole-genome sequence data to customers – but it has certainly proved the more adept of the remaining three companies at presenting large-scale genotyping data in an understandable format. The company has also been talking about sequencing for a long time
, and no doubt has been building infrastructure for the more difficult task of visualising and clarifying the function of all three billion bases in the genome. In fact, I’d be shocked if 23andMe doesn’t announce its own whole-genome sequence browser to rival Knome’s offering in the very near future.
In the meantime, though, the company has an attractive initial offer for Illumina customers: according to GenomeWeb
, up until June 2010 23andMe will offer its genotype-based service to Illumina customers at no cost
(after that “the initial analysis will cost $499 for new customers, plus a $19.95 monthly fee”). Although this analysis will mostly cover common single-base variants rather than the rare variants and structural variation that are the unique and potentially much more powerful attraction by sequencing data, a zero-cost analysis will be hard to turn down – and once customers are hooked into 23andMe’s pipeline, it will no doubt be easier to induce them to fork out a little extra for the full service.
No direct-to-consumer access
As the field evolves, “it could be in the long-run that [physician involvement] is not a requirement,” Flatley said. “We thought it’s better for us to do it this way in the beginning, to take the highest road we could find.”
I can understand this from a pragmatic point of view, especially when it comes to soothing the jittery nerves of European regulators (noting that Germany recently banned direct-to-consumer genetic testing altogether
); perhaps it is best to give society time to accept the implications of genome sequence data before pushing for broader genomic freedom. However, I’m concerned that once the medical establishment have dug their claws into the personal genomic market they will prove extremely difficult to dislodge.
Your genome on your iPhone
The Illumina service certainly doesn’t toe the conservative “genomics is serious business” line across the board – there will apparently be systems in place for customers to share their genetic information online (either anonymously or named), presumably in a similar fashion to the social networking systems established by 23andMe.
One of the cuter gimmicks unveiled by Flatley is a prototype iPhone app that will apparently allow customers to store and display their genetic variation data, and will use a biometric identification system to lock out unwanted access. To get a sense of how this would work, check out this slide-show
. This is a neat idea that will no doubt prove popular (once the price of sequencing has dropped) given the extent to which early adopters of personal genomics overlap the classic iPhone demographic.
“Ultimately , we think the data need
to be mobile connected and wind up in the cloud,” said Flatley. “You
can’t fit the entire genome onto an iPhone today, but once we’ve
calculated the Vnome (variant genome), this becomes feasible.”
Flatley dazzled the audience by
showing a preliminary concept of the iPhone personal genomes app,
produced by a developer in a mere ten days. Following fingerprint
identification, the app would present data on many different diseases
and traits. It could list by disease, drug response, or chromosome;
search by genes, share facilities with friends and family, and so on.
Finally, Flatley joked that if people spend too much time analyzing
their genome on their iPhone and the boss walks in, “all they have to
do is shake and they’ll be right back to the spreadsheet analyzing
their sales numbers.”
Illumina’s entry into the personal genome sequencing market is timely, its product is the most attractive retail whole genome sequencing offer yet to hit the market, and this announcement brings the age of the affordable genome sequence considerably closer. There is a lot to be excited about here, and I certainly have every expectation that Illumina will become a serious player in the personal genomics market.
However, perhaps I’m getting jaded with the whole personal genomics industry, but I can’t help but feel slightly underwhelmed by the details of the announcement. The price drop is subtle, the decision to farm out interpretation to other providers is pragmatic but dull, and offering the service only through clinicians is a disappointing nod to the regressive forces of medical paternalism. I suppose I naively expected something more profound, more revolutionary, rather than just another plodding step towards cheaper genome sequences.
Still, my idealism is not quite dampened yet: Illumina has done spectacularly well at unsettling well-established competition in both the genotyping and DNA sequencing markets, and I hope to see a similarly disruptive effect on the personal genomics industry. Let’s wait and see what happens next.