By Monica Heger
Organizers of the Archon Genomics X Prize — which will award $10 million to the team that first successfully sequences 100 whole human genomes in under 10 days for less than $10,000 per genome — have expanded the goal of the competition in the face of rapidly advancing sequencing technology.
While the prize and competition rules remain the same from when they were announced in 2006, the organizers acknowledge that technology has developed so rapidly that they now consider a main goal of the competition to advance genomic medicine and develop methods that are applicable for clinical sequencing, rather than just encourage the development of new sequencing technologies.
"Though the primary goal of the AGXP is to reward the most advanced sequencing technologies, we are also aware that the judging process and criteria, and the materials used for judging, will be important standards for clinical sequencing. Therefore our goal will be also be to develop protocols that can be used as industry benchmarks," a team of advisors to the Archon Genomics X-Prize competition wrote in Nature Precedings last week.
Larry Kedes, a senior advisor of the AGXP and professor of medicine at the University of Southern California said the original goal of the prize — to provide a financial incentive to drive the advancement of sequencing technology — is not as relevant as it was when the competition was first designed.
The DNA sequencing industry "is a multibillion dollar industry, so at this point I'm not sure that the X Prize is incentivizing people in the industry," Kedes told In Sequence.
Instead, "what I think it will do is set a standard. And probably for the first time, [it will] be able to give a grade, a quality assessment of the sequences being turned out by a variety of techniques," he said. "Right now there is no scoring system. Everyone claims that their sequencing systems are wonderful. But how good they are? We just don't know."
Last year, a team was formed to help draft a validation protocol for how the competition would be judged (IS 5/25/2010). Last week, the criteria were published in Nature Precedings in an effort to elicit comments from the community and draw in more competitors. Aside from Roche's 454 Life Sciences, none of the major sequencing companies have entered. After April 15, the goal is to publish a final version of the validation protocol incorporating any changes.
The 100 genomes could then continue to serve as a standard by which future sequencing technologies can be judged. Kedes said the hope is that each of the competitors would make their sequenced genomes available to the public. Then, data from the sequenced genomes of different competitors can be combined to form even more complete genomes.
"Ultimately, the 100 genomes will be [closest] to 100-percent accurate genomes ever," Kedes said.
An additional goal is to advance genomic medicine. "We believe that we need this quality for whole human genome sequencing for personalized medicine to take off," he said.
[ pagebreak ]
Since the publication of the protocol, Kedes said two additional companies have entered the competition, but as of now are still in stealth mode. He said that X Prize organizers have also been having discussions with all the major sequencing companies, including Illumina, Life Technologies, Complete Genomics, and Pacific Biosciences.
The organizers set out three main criteria to evaluate the sequenced genomes: they must be 98 percent complete, 99.999 percent accurate, and must be a full diploid genome. In addition, competitors will be judged on their ability to call SNPs, to do haplotype phasing, and identify copy number variations and structural variants, requirements included in the original guidelines.
The judges plan to evaluate the submitted genomes by a sampling approach. To that end, they will randomly sequence 1 percent of each of the 100 genomes. The corresponding section in the competitor's genome should be at least 98 percent concordant to the randomly sequenced sample.
The idea is the same for measuring accuracy. There should be less than 1 error per 100 kilobases in the section of the competitor's genome that matches the randomly selected 1 percent.
Kedes said the team decided to not include certain areas of the genome like repetitive regions and telomeres in the evaluation because those are "segments of the genome that no technology can sequence well." So, if a repetitive region or telomeric region happens to turn up in the random sampling, it will be discarded. The disadvantage of that, he said, is that if a competitor is able to sequence a repetitive region, "we'll never know."
Going forward, he said the major challenges would be to design a bioinformatics pipeline capable of analyzing the data, choosing a panel of judges, and enticing enough competitors to make the competition more valuable. The more competitors and the more sequencing technology that is used, the more valuable the sequenced genomes will be.
Since the goal of the Genomics X Prize has shifted, Complete Genomics' cofounder and chief scientific officer Rade Drmanac said the competition's requirements should also change to focus solely on what would be needed to advance genomic medicine. Right now, he said, the current competition guidelines are a "mixed bag of clinical and research conditions," and should be revised to make it relevant for clinical applications.
For instance, he said, the error rate of 1 in 100 kilobases is great for research purposes, but researchers are already achieving accuracies at this level, and "for clinical applications, we need much higher quality."
He said that there are three important areas for sequencing a genome for clinical purposes: accuracy, completeness, and cost. In terms of completeness, he said for clinical purposes it would be more valuable to decide which sequences are the most important, and judge the completeness of those sequences.
In addition, for sequencing to be done clinically, he said, it would need to cost under $1,000 per genome, not the $10,000 required by the X Prize.
He added that the requirement to do haplotype phasing is extremely important, and he liked that the X Prize team included that in the competition's requirements, especially because very few sequenced genomes also have whole-genome haplotype information.
Recently, several groups have taken steps to develop whole-genome haplotyping methods, which will be critical for bringing whole-genome sequencing to the clinic, Drmanac said, but there are still only a handful of genomes with haplotype data (IS 12/21/2010).
He agreed with Kedes that one aspect of the competition would be to produce better genomes in order to analyze sequencing technologies. "The more standards we generate, the more useful," he said. "And it's very important to evaluate the quality of a genome."
However, "we are already comparing technologies," he said. The same genomes are being sequenced by multiple platforms, and sequencing parent-child trios is useful for determining accuracy.
Drmanac said that as of now, the company does not plan to enter the competition. "For us, and many other companies that have invested hundreds of millions of dollars already, the monetary prize is not that relevant," he said.
But, he said, the competition is important for defining the direction of the field and promoting the field, and in particular, defining the important criteria for bringing sequencing to the clinic and making personalized medicine a reality.
Have topics you'd like to see covered by In Sequence? Contact the editor at mheger [at] genomeweb.com.