This fall, Affymetrix will begin rolling out its 100K SNP array to a small number of academic labs and drug makers. The product, which is slated for general launch in the second half of next year, is intended to become a linchpin in the firm’s plan to enter the SNP-analysis market.
In the short term, the move will place Affy, the world leader in DNA chip sales, at the back of a queue of highly competitive platform companies intent on holding onto a finite customer base.
“Absolutely,” said Greg Yap, Affy’s senior marketing director for DNA analysis, responding to whether the impending 100K launch, and the general introduction two months ago of the 10K array, means that the firm is gearing up to move itself formally into the SNP-analysis market. “Our goal is to be able to provide customers with tools for genetic analysis.
“These new product lines meet the growing customer demand for powerful SNP genotyping and resequencing solutions in basic research, clinical research and development, drug discovery, and pharmacogenomics,” he said in a press release in July.
Ironically, Affy’s decision to enter the DNA-analysis space mirrors a reverse move by the big SNP-genotyping companies into the gene-expression market — think Sequenom tweaking its MassArray platform to pursue expression research, and Illumina’s impending launch of a similar instrument based on its BeadArray technology. And Applied Biosystems, which has few predators, is entering both the gene-expression and genotyping markets.
In fact, Affy started life with a genotyping play. Its first commercial products in the mid-1990s were arrays for genotyping HIV, P53, and cytochrome P450 mutations, Yap said. “What we found at the time was that the opportunities for the gene-expression market sort of exploded, and that took up most of our attention through the late 1990s,” said Yap. “Now that our technology has matured a little bit, we see opportunities to leverage all of the … infrastructure … back to the original corporate goal of DNA analysis.”
The 10K and 100K SNP arrays, then, represent the firm’s re-entry into the space. In late July, the company introduced the Mapping 10K array for whole genome SNP analysis. The product, which was beta tested at around 50 labs for nine months, enables researchers to perform benchtop-based whole-genome SNP analysis using a single primer pair that can examine 10,000 SNPs in an assay of 250 ng of DNA.
Giulia Kennedy, director of genotyping research at Affy, said users of the 10K array can include 400,000 different sequences, and genotype the alleles by allele-specific hybridization. “Because we can put down so much content, so many SNPs on the array … we wanted to take advantage of that, and we didn’t want to have to do 20,000 PCRs, pull everything together, and inject it into one chip,” she told SNPtech Reporter this week.
To bypass those steps, Kennedy, whose team helped develop the 10K array, removed a fraction of the genome and amplified it with a pair of primers. “So instead of doing 10,000 PCRs, you’re doing one PCR,” she said. “A simple technician in a small academic lab with an eight-channel pipette can run hundreds of people and get 10,000 genotypes per chip in a matter of days.”
The 10K array uses data collected by the SNP Consortium, though the 100K will also contain SNPs from Affy’s spinout Perlegen Sciences, which has amassed 1.7 million SNPs. Affy licenses that database, and Perlegen uses it internally.
Kennedy said the array is scaleable, and that Affy can “boost up the amount of DNA that’s in the target 10-fold, and we can still keep very high performance.” This is the basis for the 100K, she said. “We’ve shown easily that we can go from 10K to 100K, and there’s no reason why we can’t go beyond 100K.” She added the 100K will be available as a two-chip set, with each chip comprising a different enzyme fraction.
Said Yap: “There’s clearly demand for more than 100,000 SNPs, and with the combination of photolithography and our single-primer PCR assay, that’s going to allow us to scale up and provide more information to customers.”
Affy is betting that the 100K tool — which it has dubbed Centurion — will target the full-genome-association market (the 10K product is primarily sold to the genetic-linkage market). Yap said drug makers could use the 100K to stratify trial cohorts, revive failed drugs, and expand drug labeling. “These are all areas that people have expressed interest in on the clinical side,” he said.
List price for the 10K is $400 per chip, plus $90 for reagents. Affy has yet to organize a pricing schedule for the 100K product, and the company is still finalizing the 100K beta customers, Yap said. However, David Kwiatkowski, director of the genotyping core facility at Harvard-Partners Genome Center, said he met with Affy officials last week, who told him that the informal price for the 100K is currently around $1,000 per chip.
Some early customers of the 10K appeared to be pleased with the array. Christian Lavedan, director of pharmacogenomics for Novartis, was quoted in an Affy press release describing the 10K as “the most comprehensive technology available for whole genome scans” whose assay “is amenable to high throughput genotyping analysis.”
Some competitors were unimpressed. “We don’t think this will be a big threat,” said Edvin Munk, manager of product marketing at Sequenom. With the 10K array, Affy will be covering a “very, very small percentage of the market” because the product is “very inflexible,” he said. “The 100K … is still inflexible, but you would be able to do a quick scan of the genome. And I think that probably would be a market for people who have the [GeneChip] scanners already.”
Munk said that Affy’s SNP array might be better suited to “quickly looking at the genome, and doing some rough scans, but in the end I would always predict that researchers would need some other tools to do the fine mapping.”
Some academics agreed. Kwiatkowski, whose lab has used Affy chips in the past but that currently uses the Sequenom MassArray technology along with ABI’s TaqMan, called the 10K chip “interesting” but said he does not “see it having a huge amount of use — although people will use it.”
He said the “problem” lies with Affy’s approach, which prevents researchers from fully specifying which SNPs are assessed. “So if you don’t really care, you can do the 10K chip, because the cost is pretty good for 10,000 markers. If you want to do a very low-level genome screen — like for LOH studies and tumor samples — it’s useful.” For most other applications — performing whole-genome screens or allele-association studies — Kwiatkowski said the 10K is “not so good.”
The 100K, however, “is much, much better,” he said. “How effective it will be is still uncertain because estimates as to the number of SNPs really needed [to scan the genome] properly really range from 100,000 to 500,000. And, again, though the 100K [chip] is a big step forward, these are not SNPs chosen to give you the uniform coverage that you would want if you knew a minimum set of SNPs that needed to provide that coverage.”
However, the chip’s tentative $1,000 pricetag would help drive the per-genotype cost down to 1 cent — “which is in the ballpark of where we need to be.”