It will be a while before any new sequencing technology platform delivers the $1,000 genome, but the application market appears diverse enough to sustain several technologies along the way.
That seemed to be the consensus of a roundtable discussion entitled “The $1,000 Genome: Are We There Yet?” at last week’s Cambridge Healthtech Institute conference on Next Generation Sequencing Technologies in San Diego.
“As a consumer, it is good to have competition,” said roundtable panelist Jay Shendure, a researcher in the department of genetics at Harvard Medical School.
Shendure, who helped develop polony sequencing as a graduate student in George Church’s lab, said he sees two fronts in the genome-sequencing market, depending on the application: 454’s technology and Sanger capillary electrophoresis sequencing are rivals on one side, while Illumina’s Genetic Analyzer and ABI’s SOLiD platform compete “on the other side of the line.” Different applications “have different demands,” he said, and therefore may benefit from different platforms.
For example, while genome assemblies are more easily done with long reads and mate pairs, he said, extremely accurate consensus sequences of “less than one error per gigabase” will be critical for medical resequencing projects.
Bob Strausberg, deputy director of the J. Craig Venter Institute, agreed that the type of technology to be used “depends on the area of interest.” The Venter Institute has 454’s new FLX system in place and is working on a sequencing project in collaboration with Illumina, he said. But it is also still making use of its large farm of ABI 3730’s.
Especially in large metagenomics projects, where “you are looking at a pot of genes” that are hard to assemble, only the long reads from Sanger sequencing allow researchers to identify protein species, he said.
In cancer sequencing, however, the “digital” readout of the new technologies trumps the “analog” readout of the Sanger method, which generates an average signal from many individual DNA molecules, and the new platforms allow researchers to find mutations “we cannot see with Sanger sequencing.”
Michael Rhodes, ABI’s applications manager for high-throughput discovery, also sees continued use for Sanger sequencing, especially for low-throughput, fast-turnaround needs, such as “one amplicon, in one individual, this afternoon.” Last year, ABI saw growth in this segment of the business, he said.
But according to Michael Egholm, 454’s vice president of molecular biology, the days of Sanger sequencing are numbered. In the future, 454’s technology will replace Sanger sequencing “for most applications,” he believes. “We are only at the beginning.”
Debating the $1,000 Human Genome
Panelists also differed on how much it will cost to sequence a human genome by 2010. According to Egholm, it currently costs around $1 million to do that kind of project on 454’s instrument, and “people who can afford to do that are interested,” he said.
Egholm predicted that cost will drop to between $10,000 and $100,000 by 2010 “if not before,” which is “going to be very attractive” to many.
454 is in the process of finishing “Project Jim”, the genome sequence of Jim Watson, which the company plans to present at a meeting in May (see In Sequence 3/13/2007).
Harvard’s Shendure predicted it will cost around $25,000 to sequence a human genome by 2010.
These predictions are less optimistic than the results of a recent CHI survey, which found that one-third of approximately 100 respondents who CHI identified as “active in or following the field” said they believe the cost will come down to between $10,000 and $25,000 by 2010. Around half said it will be lower than that.
How valuable will the $1,000 genome be? According to Strausberg, researchers should first sequence a “considerable number” of human genomes before deciding how informative whole-genome sequence will be. He mentioned that the Venter Institute is currently re-sequencing Craig Venter’s genome.
Already today, genotyping arrays “can tell you most of the common SNPs for $1,000,” Shendure pointed out, and it remains to be seen how much meaningful information whole-genome sequencing can add to that.
“Sequencing Jim [Watson] in itself is kind of uninteresting,” he said. “Right now, there is nothing we can do” with much of the sequence information, he added, and new statistical approaches are required to interpret, for example, rare alleles.
On the computational side, storing and analyzing terabytes of sequence data will be a challenge, the panelists agreed, but whether the raw data needs to be stored indefinitely remains under debate.
How valuable will the $1,000 genome be? Today, genotyping arrays “can tell you most of the common SNPs for $1,000,” and it remains to be seen how much meaningful information whole-genome sequencing can add to that.
Rhodes estimated that storing 6 terabytes of raw data per week — data ABI’s SOLiD platform could generate in a single run — would cost on the order of $1 million per year, and there is no need to keep this image data forever.
454’s Egholm, on the other hand, said his company has “stored all data so far,” and that for future clinical applications it might be necessary to keep the raw data.
Asked about the next-next-generation technologies, in particular nanopore sequencing, the panelists showed caution. “If it works, it could change the world,” said Rhodes.
Shendure pointed out that the idea of nanopore sequencing has been around since the mid-1980s. With the current investment by the National Institutes of Health in the “best ideas,” he said, “we will know within one or two years” what is possible. “If the investment does not pan out, it will be a long time,” he said.
What would the panelists do with a $1,000 genome technology? For Shendure and Strausberg, applications in cancer genomics would be most interesting. Later at the meeting, Shendure presented new techniques for selecting portions of the human genome for sequencing, such as the exons, and Strausberg presented an analysis of glioblastomas using Sanger and 454 technology (see In Sequence’s sister publication, GenomeWeb Daily News, 7/24/2006). Egholm proposed to “take on some common disease and then really nail it” with the sequencing data.
Rhodes, meantime, said he would like to “sequence everything that’s on the Earth.”
The panel discussion was part of a two-day meeting that attracted more than 80 participants. CHI held the meeting for the first time last week and plans to repeat it every year.
The meeting featured a variety of approaches to and applications for high-throughput sequencing, ranging from 454’s and ABI’s next-gen platforms to methods that are still under development (for an example, see article in this issue).
Illumina, which is based in San Diego, was conspicuously absent from the program and exhibit hall, but its presence was felt in discussions throughout the meeting. Conference organizers said Illumina was invited to participate, and Illumina representatives did not respond to requests for comment.