"Today we are learning the language with which God created life." President Bill Clinton made this remark on the White House lawn on June 2000 to recognize the decoding of the first human genome. As much as anything else, rapid DNA sequencing technology created in large part by geneticist Craig Venter and his colleagues galvanized the research community into finishing the project faster than originally expected. More than 11 years later, however, gene sequencing technology has failed to deliver on its promise to revolutionize preventative medicine, and Venter is not happy about it.

The idea was that gene sequencing would become so cheap—on the order of $1,000—that ordinary people could afford to have their individual genomes sequenced, which their family doctors would use to diagnose their predisposition for disease. Costs have fallen to about $4000, but the bigger problem is that results are often rife with errors. “If [the technology] is going to achieve the level of really impacting medicine the way I’ve always envisioned that it could, it has to become far more accurate.”

To goose developers along, Venter—along with the X-Prize Foundation and Medco Health Systems, a health care firm—has put a bounty on the achievement: $10 million to the first team that can meet a standard for accuracy that Venter calls “medical grade.” The task is to sequence the genomes of 100 centenarians, providing a baseline to which other genomes can be compared. Venter talked with Scientific American by phone about the award and his hopes for new technologies. Excerpts:

Scientific American: How did the idea for this prize come about?

Venter: This prize started out as a half million dollar prize out of the Venter Institute, in part after I sequenced the first version of the human genome. All the analysts and pundits were saying that genome sequencing was dead, there’s no use for it anymore—the human genome has been sequenced. To change that attitude, we started a prize to encourage development of new technology, to get [costs] down to a $1000 or less [per genome].

I was contacted by Larry Page and Peter Diamandis of the X Prize to see if we were interested in merging the Venter Prize with the X Prize. We did that and upped the ante to $10 million, which is significant.

Why is a prize necessary?

The technology is changing pretty rapidly, which is a good thing. But right now there’s no technology out there that meets the standards that we’ve set. If genome sequencing is going to have true medical impact, it needs to get up to [a higher] diagnostic quality level. And we’re a long way from that. So this announcement is a whole new set of guidelines and standards. We’re also working with the FDA trying to have our standards become their standards for sequencing. The fact that Medco has come in as a major sponsor shows that major medical enterprises are starting to recognize the future medical impact of this technology.

So the idea is to give the technology a little extra push?

It’s helpful to drive the technology forward. The cost is coming down, but one thing that hasn’t substantially changed is the completeness and accuracy. You can buy any two company’s machines right now, and if we sequencing a genome with two different technologies we’d get two different answers. That’s not good for diagnostic sequencing. Each technology has its own inherent errors. If [the technology] is going to achieve the level of really impacting medicine the way I’ve always envisioned that it could, it has to become far more accurate.

Don’t companies like 23andMe already do diagnostic sequencing?

They don’t do true sequencing, they just measure a tiny portion of the genome. So we’re trying to drive things to actually measure the entire human genome and do it accurately.

Why do you need the whole genome to make a good diagnosis?

A decade go that [the BRCA1 gene] was thought to be a predictive gene for breast cancer. Doctors would tell women if get they had that gene they had a 99 percent chance of getting breast cancer. It turns out that wasn’t true. They extrapolated from genetic studies of unique populations to the general population. But people have other genes and other traits that protect them from getting breast cancer.

We’re still at an early stage in genetics. I was one of the first people to have his whole genome sequenced, but we still can’t tell very much about me from my genome. We need to have definitive phenotypes for purposes of comparison.

Why are you specifying that contestants sequence centenarians?

These people are the Michael Jordan’s of longevity? They’ve already won that contest. We wanted to do see if there’s something unique they have in common with them—wellness genes that would protect you from cancer if you were genetically predisposed to getting it.

You’d match the genomes to the centenarians without necessarily understanding what you were comparing?

You’re not going to get a definitive answer just by looking at 100 genomes [of ordinary people]. Centenarians is a more interesting group than most because it emphasizes the wellness aspect rather than the illness aspect. And that’s what we want to emphasize, what’s going to get us over disease and to live to over a hundred if we want to.

If there’s so much promise here, why isn’t everyone already working as hard as they can on this? Why do we need a prize?

Well, there’s hundreds of mllions of dollars being invested in this space right now. The technology side is pretty miraculous. In a decade, going from a $3 billion federal program to where you can call a couple of companies to get your genome sequenced for $4000—that’s a pretty dramatic change, right? But that’s still probably order of magnitudes too high to make it generally applicable to large numbers of people. We need to get large numbers to make the [genomic] information meaningful.

People have focused on technology and cheapness, but they haven’t focused so much on accuracy, in part because there hasn’t been demand for it, so they’ve been comfortable with one or two or three percent error rate. If you’re interested in being diagnosed, you’re not going to be comfortable with that kind of error.

What happens when we get an accurate sequencer—that will it mean?

We’re defining the standards. This prize becomes the truth serum for all the claims from all these companies and all these technologies. Without an independent test, there is no ability to sort out the claims, and there are some pretty wild ones out there. The hope is there will be one or several winners out there. For guys and gals working in their garages, [the $10 million prize] is a strong incentive.

Do you think the technology is out there in someone’s garage?

It’s not in any of the production machines I’ve seen. But there are lots of new techs coming. I don’t think any of the [current] machines could win this contest.

I know it can be done in a garage. I had a conversation yesterday with a couple of guys that have developed a radical new technology nobody’s thought of before, and they’ve put it together with a fairly simple budget. We’re early in this field. Biology is a young field.

Why would this technology be doable now and not 10 years ago, when you were working at Celera?

Look what’s changed in the last 10 years. In 1999, at Celera, we had a giant computer that took up an entire room and could perform one-and-a-half trillion [operations per second]. Now you can get a little card that you plug into your computer desktop to turn it into a tera-op machine. [Computation is very important in] genome sequencing, in fact it’s the hidden part of this that is critical, and that alone has changed, following Moore’s law. Somebody starting out today has orders of magnitude more information to start form than someone 10 or 15 years ago did.