For Affymetrix, the settlement with Incyte caps off what has been been a remarkable comeback in the second half of 2001.
“This has been an incredible year for Affymetrix,” said company spokesperson Anne Bowdidge. “It has been an challenging year as well.”
The Ides of March came early for the Santa Clara, Calif. microarray manufacturer when the company officially disclosed March 8 that its U74 Murine arrays contained defective sequence due to an error in reading the direction of sequence in the UniGene database.
While the company quickly quelled protests coming from the microrrary community by promising immediately to replace the arrays one-for-one within six weeks, the disclosure was a major setback for a company premised on providing reliable research tools.
Then in June, the company issued a warning that its second-quarter revenues would total between $6 and $12 million less than the $56 million expected, largely due to the murine array debacle. Not only had the company incurred an additional $5 million in direct replacement costs for its defective U74 murine arrays, the company also incurred a bigger loss in revenues due to customers’ hesitancy in ordering Affymetrix chips. On the spotted array side, the company’s arrayers and equipment were not doing as well as expected — a matter company president Sue Siegel attributed to the shift from do-it-yourself arraying to pre-fabricated arrays.
But after reaching this nadir in the summer, the company rebounded. The second quarter revenues were not as bad as feared, the third quarter revenues rose to $55.4 million, exceeding Wall Street’s expectations. Also, the company saw an upswing in sales of GeneChips, signaling that the sector had regained confidence in its products. Reflecting this rebound, analysts that had downgraded the company’s stock in June reinstated their “buy” ratings and the company’s stock price surged.
Meanwhile, in the vulnerable periodfollowing the disclosure of defective chips, the company had begun to show a new attitude of openness to critics. The two biggest criticisms of the GeneChip platform, other than its price, included the company’s failure to disclose the sequence of the oligonucleotides used as probes, and its array analysis algorithm, which some statisticians said relied on a false assumption that microarray data fit a normal distribution curve.
Answering these critics, Affymetrix released its new website, NetAffx, where potential customers could gain access to information about the genes corresponding to the sequence on the arrays. Then in late August, the company announced it would replace its empirical algorithm with a new statistically-based algorithm in version 5.0 of its software. The company has since explained the algorithm at numerous conferences and included explanations of the algorithm on its website. Finally, company officials told BioArray News in late October that the company planned to reveal the sequences on its oligonucleotide probes early this year.
This new openness and willingness to respond to customer feedback does not just reflect a policy of goodwill, but more fundamentally the company’s increasingly comfortable position in a rearranged microarray sector. When Corning and Incyte both agreed to pull out of the microarray business in October, Affymetrix lost two challengers in the high-density array field who both had formidable capital and manufacturing capabilities. This left Motorola, Agilent, and a crowded waiting room of startups.
Neither Motorola nor Agilent has yet been able to put a dent into Affymetrix’s market share, even though both have been at it for over two years. And as Affy grows bigger, with a global staff of over 900 and expanding facilities in the UK and California, these two competitors — both of which are attached to parent companies that have suffered substantial financial blows in the past year — may be forced into niches or off of the microarray radar altogether if they do not step up their marketing efforts and find new ways to convince researchers their wares are superior to the GeneChip, a product that has already won the mindshare of the scientific community.
Affable, But Still Facing Obstacles
But this doesn’t mean Affymetrix can glide passively toward profitability, which it expects to reach in the second half of the year. The company still faces two obstacles before it wins over the hearts and minds of the remaining microarray market: price, and customizability. The price issue could be addressed through the development of an ultra-high density array that offers the entire genome on one chip and sells for the same amount as the current package of five chips with 60,000 genes and ESTs. Affymetrix has already stated that it intends to move toward this goal, but like any company will set its prices according to what the market will bear.
The other issue, customizability, is more tough to solve than price. Currently, many scientists use Affymetrix arrays for that first-pass of exploratory research, then once they have located a set of 50 to 200 genes of interest, spot down their own custom arrays for repeated experiments on the handful of genes. When Millennium Pharmaceuticals signed on this October as Affymetrix’s largest customer, switching over from a do-it-yourself platform, company officials expressed a desire to retain some spotting capabilities for this sort of high-throughput testing using a small number of genes and thousands of compounds or conditions.
The CustomExpress arrays, which Affymetirx introduced with the new NetAffx website in July, go only part way toward addressing this need. They are only available in large quantities due to the need to make a photolithographic mask for each kind of array, and take a minimum of four weeks to manufacture and deliver. Spotted arrays, by contrast can be made in small batches cheaply and revised overnight.
There is, however, an obvious experimental advantage to Affymetrix customers in using the same oligo sequences they used in the initial high-density experiments: Given that genes have multiple splice variants, researchers using Affymetrix oligo probes in the second stage could more safely assume that they were probing for the same gene or gene variant they found expressed in their initial experiment, while those using cDNAs spotted down might get a different result. They could also avoid the hassles of verifying clones through RT-PCR and the chip-to-chip variabilities inherent in spotted array platforms.
If Affymetrix can develop — or acquire — the technology to make its custom arrays more flexible, and faster to produce, then this could be the death knell for do-it-yourself systems. Then, the emerging GeneChip giant will have only one major adversary to worry about: a well-run startup with technology that will make its microarrays obsolete.