The market for expression arrays is often characterized as mature, a slow-growth sector where Affymetrix and Agilent Technologies dominate, and the only options are to get out — like GE Healthcare has done — or to work hard to steal market share from the juggernauts, which is the path Illumina has chosen.
However, in recent weeks Illumina has told investors that the sluggish gene-expression plateau is only temporary, and that the company has the remedy to expand it from its estimated $800 million market size with double-digit growth: its Genome Analyzer sequencer, which it obtained through its acquisition of Solexa in January
On two occasions in the last three weeks, Illumina CEO Jay Flatley told investors that the market for traditional, fluorescence-based “analog” arrays will eventually be eclipsed by demand for “digital gene expression” studies performed on the firm’s Genome Analyzer next-generation sequencing instrument.
At the Bank of America Healthcare Conference in Las Vegas three weeks ago, Flatley told investors that “array-based gene expression is going to die out pretty fast” once digital gene expression using the Analyzer becomes popular (see BAN 6/5/2007). And last week at the Goldman Sachs Healthcare Conference in Dana Point, Calif., he reiterated his prediction for the future of analog arrays that companies like Affymetrix, Agilent, and Illumina currently sell.
“We had a researcher who we are very familiar with and who is a very fundamental person in this area demo the system and his statement after that demo was that DNA chips are very soon going to be dead, be passé,” Flatley said in Dana Point.
“Our hope is that [the Analyzer] can reinvigorate the gene-expression market and get it to be a significant grower again,” he said. “That’s ahead of us, we have to show that, but our hope is that people will embrace this new paradigm of expression and get this market back to growth rates of 10 or 20 percent per year,” Flatley added.
No Clear Timeline for Market Shift
Some array users share Illumina’s optimism about digital gene expression, but the big question remains when that will happen. According to most, the demise of the analog array is a long way off, even if it is inevitable.
According to Flatley, Illumina is not even “sure exactly how fast that is going to happen,” but he is certain that the benefits of the approach, which allows users to “digitally count transcripts in the genome and to do that without knowing the sequence” is “a new way to do expression and to do it simultaneously across the genome” that will ultimately drive adoption.
Flatley said that the firm already has users in the field using the Genome Analyzer for digital gene expression. He did not respond to an e-mail seeking additional details by press time and no papers have been published describing the use of the Genome Analyzer for digital gene expression to date.
During the Goldman Sachs conference, Flatley explained why Illumina sees the sequencer as a bridge over existing expression arrays, including its own.
“From a performance perspective, in an array you are measuring the intensity of a fluorescence signal,” he said. “And so you are always looking, let’s say in a scanner, in trying to distinguish between 600 counts of fluorescent light and 640 counts of fluorescent light, and it’s always hard to rely on that difference. At some point you have a noise floor where you really shouldn’t differentiate below a certain level,” he said.
“This creates challenges in detecting rare transcripts where scientists think there is an abundance of information that may be related to disease, that we just don’t detect today,” he added.
In comparison, Flatley said that sequencing is a “very high-throughput way to do [gene expression]. It’s digital so there is no noise; we have the ability to look across the entire genome, so you don’t look just in the genes where most of the array technology has been focused to date.”
He added that “because of that you don’t need to know the sequence in advance, so it gives you the ability to count these expression transcripts in organisms that have not been sequenced.”
The Digital Bandwagon
Flatley is not alone in imagining a world without analog arrays. Other companies that sell next-generation sequencers have made similar claims of their gene expression capabilities. For example, Helicos CEO Stanley Lapidus told BioArray News last year that the firm’s technology is capable of "a range of things that are not necessarily possible with today's arrays.”
Lapidus said at the time that Helicos’ goal “isn't to compete with Affy for a price point because that makes no sense. Our goal is to make available new capabilities that can't be done on array-based technology at any price" (see BAN 6/13/2006).
Some long-term array users already believe that digital gene expression will replace analog gene expression. Shrikant Mane, who heads Yale University's Affymetrix GeneChip core and is responsible for all other array platforms offered at Yale, told BioArray News in an e-mail this week that “digital gene expression will in near future replace current methods of hybridization-based analog gene expression.”
According to Mane, “digital gene expression involves less labor and may introduce [fewer] artifacts and [less] noise. As the next-generation sequencing technologies evolve, digital gene expression profiling will offer [a] more cost-effective alternative.”
Roderick Jensen, the director of the Virginia Bioinformatics Institute's genomics core laboratory, told BioArray News this week in an e-mail that he, too, believes that “measuring gene expression by direct sequencing will soon be very competitive with traditional microarray assays.”
“In addition to giving a direct readout of how many copies of each gene are present in the sample, you also get information about splice variants, gene fusions, and single nucleotide polymorphisms in the expressed genes,” he wrote. “You also get information about other transcribed sequences, such as miRNAs, and maybe things we don't even know about yet. As the prices for this technology drop in pursuit of the $1,000 human genome, this technology will eventually overtake the DNA microarrays,” Jensen wrote.
Though the triumph of digital gene expression seems inevitable to some, even Flatley conceded last week that it will be hard to tell researchers to surrender their mountains of analog-array-generated data for a method that has yet to produce significant volumes of literature.
“The pushback might be that the present data and the digital data are not comparable and that you might have to start all over again with an absolute measure. But once you get the digital data, digital data is absolutely more comparable between tissues and over time, it is much more long-term useful information,” he said.
“In terms of cost, to do what you can do on arrays right now, we can do it at about equivalent cost on a sequencer right now. If you wanted to do the entire transcriptome, that is still going to be more expensive than what you can do on an array, but over time we believe the price will come down,” Flatley added.
Along with the issues of existing data and cost, there’s also the fact that Applied Biosystems acquired a sequencer through its purchase of Agencourt last year, but neither of the expression market leaders, Affy and Agilent, have followed suit, which could indicate that they do not see sequencing as a competitive a technology to analog arrays as Illumina does.
Affy has so far remained reluctant to the idea of acquiring its own sequencer to compete against Illumina and others. Former Affy Chief Financial Officer Greg Schiffman said last November that Affy sees the next-gen sequencing market as too immature to warrant the investment.
“I think that while these technologies are exciting, this is not a near-term revenue driver for us anyway. We think it’s just a little bit early to be placing our bets at this point,” he said at the time (see BAN 11/21/2006).
Affy and Agilent did not provide comment by press time.
“Our hope is that this technology can reinvigorate the gene expression market and get it to be a significant grower again. … Our hope is that people will embrace this new paradigm of expression and get this market back to growth rates of 10 or 20 percent per year.”
Finally, some array users believe in the potential of digital expression, but they aren’t holding their breath for the day when analog arrays are a thing of the past. Bing Ren, assistant professor of cellular and molecular medicine and chromatin immunoprecipitation (ChIP)-on-chip pioneer, told BioArray News this week in an e-mail that “it is still early to tell” how digital gene expression will perform, though “some recent publications have put the two platforms — sequencing vs. microarrays — on par to each other.”
“In the long term, if price of sequencing and quality of reads keep improv[ing] as they have been in the recent years, then sequencing is indeed likely to replace microarrays for ChIP-chip,” he wrote.
Andres Metspalu, head of the department of biotechnology at the University of Tartu in Estonia told BioArray News in an e-mail this week that he is not convinced that digital gene expression will hasten the end of microarrays as we know them.
“I believe that array platforms will stay for many years,” he wrote. “New sequencing platforms will have strong input in the next discovery phase in genomics, but for the common variants current arrays are good for a long time,” he wrote.
“Nothing is ideal and multiple platforms and methods will be used in [the] future as well as today,” he added. Metspalu wrote that if there is a shift in array usage, it will be towards the diagnostics arena, while sequencing might become more valued in the research arena.
“Their main focus might be shifted more from research to diagnostics, not only for disease, in coming years,” he wrote, and the “$1,000 test for the genome will not replace the arrays, but will complement them.”
VBI’s Jensen agreed that, rather than sequencing replacing arrays, the use of both platforms is likely to become the norm as sequencers become more available to academic researchers.
“In the short run there are many interesting opportunities to use the two technologies together,” he wrote this week. For example, Jensen said that analog and digital expression could be paired to “identify new splice variants and SNPs or other transcribed sequences in disease samples and then design custom microarrays with probes for these novel biomarkers for high-throughput