When Motorola hung up on its hopes for the high-density array market last week, selling its CodeLink high-density microarray business to Amersham Biosciences for $20 million, this created quite a bit of static in the microarray sector: Amersham is now poised to grab a piece of the market from Affymetrix, which just strutted through a third straight quarter of growth.
“We have taken control of this business because we believe that we can take a substantial market share over the next one to two years,” said Andrew Carr, president of Amersham Biosciences, in a conference call to discuss the acquisition. The company is looking to build up the market for the CodeLink arrays from the existing $5 million customer base into “the order of 20 to 25 percent,” of the market or $150 million, Carr said.
Amersham is not just bluffing. The UK-based lord of the life sciences, which has nearly $1 billion in annual sales, 50,000 customers, and 4,400 employees around the world, obviously dwarfs the Santa Clara oligo oligarch in its global reach and market power — a fact of which Affymetrix execs are all too aware. After all, president Sue Siegel and chief commercial officer Trevor Nicholls are both Amersham alums, and up until now, Amersham''s Japanese subsidiary has distributed Affy chips in that country. Additionally, Amersham has an established presence in the microarray field, with its CyDye products, and its Lucidea spotters, hybridization stations, software, and QC reagent kit. The company also distributes Axon GenePix scanners, ArrayVision imaging software, and Spotfire’s DecisionSite platform.
Amersham plans to move the microarray business to Tempe, Ariz., where Motorola already manufactures the arrays, and retain 80 of Motorola’s 214 employees.
“I think Amersham is up to the challenge,” said Nancy Schmelkin, who has been product manager for the CodeLink arrays. “They’ve got a smart sales force out there, and a certain understanding of how to play in the life sciences arena. They also have the wherewithal to play in the life sciences arena, which is the only thing we couldn’t bring to the party as a non-life-sciences player. It would be a rare marketplace where one company dominates forever.”
Indeed, if global distribution networks trump innovation, Affymetrix could see its market share quickly erode. Carr is betting that customers will not only want the ultra-high-density whole-genome chips, Affymetrix’s strong suit, but that there will be “a very big market for theme chips” with fewer data points.
So far, a number of companies have tried to strike Affymetrix on this low-density Achilles’ heel, but without much success. The company has responded with its own 8,000-gene focus arrays recently, but has found that these chips are not as popular as its whole-genome arrays.
But Amersham could also pierce Affy’s armor at another weak point: the custom array area. While Affymetrix has offered both CustomExpress arrays, which can be designed through its NetAffx web site, and larger custom orders, the company’s in situ photolithographic process requires these chips to be ordered in batches and produced over a period of weeks. The CodeLink chips, by contrast, consist of oligos spotted down on treated glass slides with a non-contact spotter, and can be modified with every batch. Carr is aware of this potential advantage, noting the “important differences in the setup costs” for the CodeLink arrays. “We can manufacture these slides with appropriate new content onto the surfaces relatively quickly,” he said.
We’re Better. No, We’re Better.
Carr and Sir William Castell, chief executive of Amersham Biosciences’ parent company Amersham plc, have both said they think the CodeLink arrays are higher quality than Affymetrix arrays.
“The Motorola CodeLink arrays ...give more usable data points than any other pre-arrayed slides,” Castell said. Since the CodeLink slides include a maximum of 10,000 oligos while Affymetrix’s U133 includes 22,500 probe sets, the implication in this statement was that more than half of the Affymetrix probe sets don’t work.
Carr later elaborated on this theme, saying that with the CodeLink arrays “you know exactly what you put down. You are not dependent on a process on the surface of the slide that may give you or may not give you a particular molecule” — a dig at the photolithographic process that Affymetrix uses. “There are significant advantages in terms of QC security,” he added, especially if the product were used in the diagnostic arena.
While no large-scale independent comparisons of Motorola vs. Affymetrix arrays have been published, anecdotal reports from users have found that both offer reproducible results. Winston Kuo, a postdoc at Harvard Medical School, has found in an initial comparison of Affymetrix vs. Motorola CodeLink arrays on two breast cancer cell lines that the CodeLink arrays did have a higher specificity, but that both seemed to perform well. (See BioArray News, 6/21/02).
Motorola had a 20K human array in late-stage development, according to Schmelkin, and this array could compete with Affymetrix in the high-density arena. While Amersham spokeswoman Marcy Saack said the company has not yet outlined specific development strategies for this product, she did say it will go ahead with plans to bring out a mouse UniSet expression 10K array and a human genome linkage SNP chip with probes for 1,000 SNPs.
Over the next couple of years, Amersham has further plans to take the CodeLink arrays into the proteomics arena. Carr said that the slide surface chemistry, which combines acryl amide derivatives and other chemical modifications, is not only suitable for oligos, but also for proteins. “We have a lot of surface chemistry expertise that would allow us to put either universal binding agents or antibodies down onto this surface,” he said, “although our clear first priority is to extend our reach into the gene expression and SNP market.”