Dec. 17, 2008 at 10:50 PM ET
Gary Parker / Complete Genomics
A production staff member works in Complete Genomics' sequencing instrument
development lab. Improved tools are bringing down the cost of genetic analysis.
How much does it cost to decode your genome? Last year, the going rate was $1 million. Now prices are plunging - and as a result, the prospects for personalized medicine and other genetic innovations are rising.
To get a sense of how deeply prices are plunging, let's start with the whopping price tag of $3 billion for the Human Genome Project, which produced a composite readout of the DNA code from many donors by 2003.
It took a few years more to publish the first complete genome for a single human - specifically, genetic entrepreneur Craig Venter - at an estimated cost of $70 million to $100 million. Nobel-winning biologist James Watson's genome was also done up last year at a cost of roughly $1 million.
In the past year, genome-decoding has gone commercial - almost to the point of sparking a price war. A 2-year-old company called 23andMe is offering an analysis of 600,000 key DNA markers for $399 (marked down from $999). Other companies - including deCODE and Navigenics - are in the marker-analysis business as well, with services listed at prices ranging from less than $1,000 to $2,500.
You can get your entire 6-billion-base-pair genome decoded by a 1-year-old company called Knome at a cost of $100,000 (slashed from $350,000). And now Complete Genomics is gearing up to provide whole-genome analysis for $5,000 a pop.
"We are a wholesaler of complete human genomes to a variety of markets," said Cliff Reid, Complete Genomic's chairman, president and chief executive officer.
Jorge Conde, chief executive officer for Knome, noted that the target price for decoding an entire human genome is $1,000. "At or below $1,000 for a genome, this is a technology that will have a significant benefit to individuals and will be widely accessible," he said.
What's it worth?
What's the point of having your genome done? All these companies are trying to offer data that can be correlated with genetic predisposition to health conditions ranging from Alzheimer's disease and heart problems to baldness and flaky earwax.
Conde cited the example of a client whose genome pointed up a rare genetic variant for eye disorders. Even though the client had perfect vision and was experiencing no problems, he followed up on the genetic report and consulted an eye doctor.
"According to him, the physician found the very, very early stages of glaucoma, and he was able to get it treated," Conde said.
It's easy to oversell those kinds of stories, and this year the New England Journal of Medicine counseled doctors and patients to be skeptical of such tests. Even the gene-deciphering companies themselves emphasize that their results shouldn't be taken as medical diagnoses.
The pros and cons of lending your data
The biggest benefit is likely to come in the form of research that points physicians toward new avenues for fighting disease. The key is to have large databases of genotypes (genetic markers) that can be correlated with phenotypes (physical manifestations of those markers). The larger the database, the better, Reid said.
"One genome isn't really very useful all by itself," he told me. Lower cost translates into a bigger market, which in turn yields more valuable information.
Reid's goal for Complete Genomics is to decode 1 million genomes in the next five years. "The nice way to think of 1 million genomes is, that's 1,000 people in [each of] 1,000 disease studies," he said.
But that raises questions about information privacy and security - issues that will have to be resolved by the time the $1,000 genome becomes a reality. Just this year, Congress overwhelmingly approved a law against genetic discrimination - a move that generated plenty of controversy.
Knome tries to address the privacy issue by providing its $100,000-a-genome clients with an 8-gigabyte USB drive bearing the encrypted genetic code. The company itself doesn't store the data files - although they're archived by a trusted third party in case you lose that thumb drive. If researchers want access to your genetic data or medical history, Knome puts them in touch with you, and you tell them how much information you're willing to share.
Wholesale vs. retail
Complete Genomics' wholesale approach stands in contrast to Knome's retail approach. Reid's company merely churns out the data, using DNA arrays and code-readers that he says are 10 times more cost-effective than the industry standard, and then delivers the data to its customers. Those customers may include researchers looking for links between genetic markers and medical conditions, or pharmaceutical companies trying to figure out how different drugs affect genetically distinct populations, or private individuals trying to get a fix on their future health.
Although the $1,000 genome is an easy-to-remember benchmark, the appropriate price points will likely depend on how the genome is used, Reid said. On one hand, the price point might be $500 or less for folks who are just trying to figure out where their families came from, based on genetic markers linked to geographic origins. On the other hand, "I think there are going to be many applications where $5,000 is going to be a fabulous price," Reid said.
One application will be a genome-based survey of Luxembourg's population - supported by that country's government, modeled after Iceland's deCODE project and undertaken by Complete Genomics in cooperation with the Seattle-based Institute for Systems Biology.
Reid said 100 of the 1,000 genome sequencings that his company is planning to do in the next year will be devoted to the Luxembourg project, as will 2,000 of the 20,000 sequencings projected for 2010. The partners in the project expect their work to result in scientific papers that will help plot a course for personalized medicine.
Based on different genetic profiles, medical treatments eventually could be tailored to "get the right drug to the right person, at the right dose at the right time," Reid said.
Faster than Moore's Law
Knome's Conde said further advances in gene-sequencing technology, combined with higher volumes, could push down the price point faster than expected. "People talk about Moore's Law. If you look at the chart and do the math, this is happening significantly faster than Moore's Law," he said.
"The mystical goal of the $1,000 genome is likely achievable," Conde said. "It's something we'll see in the next decade, and potentially sooner than that."
Will the $1,000 genome make services that analyze only key portions of the genome obsolete? Rachel Cohen, manager of communications for 23andMe, said it won't. Different levels of analysis will serve for different genetic goals - for example, finding out whether you're predisposed to develop a rare disease vs. finding out where your ancestors came from. In fact, 23andMe might get into the whole-genome business itself.
"When we're satisfied with the data quality, that's something that we'll incorporate into the service," Cohen said.
Update for 3 a.m. ET Dec. 18: Venture capitalist/journalist Esther Dyson is an investor in 23andMe as well as a member of the Personal Genome Project's PGP-10 - a group of prominent people who are sharing their genetic and medical records with the public. I sent her an e-mail asking for her perspective on the genome market, and she wrote back from Russia (where she's currently going through cosmonaut training).
Has her perspective on personal genomics changed, now that she's had a chance to delve into her own genome? "No, there's not much to delve into, actually, other than sharing with other people," she replied. "We don't know enough yet for it to be meaningful - unless you are fascinated with the precise chances of a particular condition.
"I found comparing family members - even though that is also just percentages - to be far more interesting ... in part because those people are concrete, whereas the conditions are just general things."
She also sent along the transcript of an interview she gave to the German magazine Neue Gegenwart. Here are a couple of the magazine's questions and her answers on genetic matters:
"Q: How do you make sure that the 23andMe customer is prepared for his results? The results might change the customer's attitude to life (and 23andMe seems to be a perfect tool for hypochondriacs).
"A: We can't make sure. ... But the more people out there who have used it, the more press there is about it, the better educated more people will be. You have to start somewhere. Meanwhile, we try to be very clear in all our communications about what kind of info we provide and what it does (and does not) mean. In a sense, we are an educational tool, using people's own data as content. ...
"Q: Do you think that one day it will become normal to analyze (and publish) personal genetic information on the Web? (Why?) Insurance companies might be interested in the genetic information (and not only in your lifestyle).
"A: As you know, I have published mine already. (And for what it's worth, I offered the data to my health insurance company, but they didn't want it!) But yes, I think it's wonderful to have data on the Web to be analyzed by independent researchers, each of them with different hypotheses and analytical tools. However, most genetic data will probably remain anonymized, just as most health information will. It's not important to know the identities in doing most research.
"Q: Can't that lead to a situation where it seems suspicious if you refuse publishing your genetic information on the Web?
"A: I doubt it. Why should everyone publish it on the Web? They don't publish their financial info on the Web either."
To learn more about the promise (and pitfalls) of personal genomics, check out this report from Nature Biotechnology, and this one from The New York Times. Here's still more about genomics from Cosmic Log: