Outside the front entrance to the Affymetrix manufacturing center in Sacramento, Calif., about a half a football field away, is a red fence. It marks the perimeter of the company’s property. Like the West, there is plenty of room for expansion.
So too with the company’s product, the GeneChip brand microarray. The company has just begun to manufacture its glass microarray wafers at a feature size of 18 microns. The first product manufactured at this feature size is the company’s human genome array chip introduced earlier in the year. The next product will be a rat chip, with a single glass wafer being sliced into 64 arrays.
Company officials say that over the next few years they will travel down a path to 10-micron format chips that each can query the whole genome.
The company employs 98 operators who make arrays in a 50,000 square feet factory space, shining ultraviolet lights on array-synthesis machines in glass cases set on five-ton granite blocks. The factory, which operates 24-7, churned out 98,000 chips in the third quarter, and sometime early in 2003, the millionth chip will come off the assembly line.
The arrays begin as a 5x5 inch sheet of pure quartz glass that is sliced into 49 thumbnail sized chips, each of which sells for approximately $600. Each chip has 10,000 features of 18-micron size, covered with 25-mer oligos.
At each incremental step down the feature-size ladder, more and more features can be added. A diamond-covered saw, which costs $50 and can make 400 passes before going dull, can slice a wafer into 64, 81, 400, 1,600, or even more chips.
Affymetrix sees its manufacturing setup as a competitive advantage. It can slice and dice its wafers into smaller pieces, albeit at a reduction in content, at the same feature size. Or, it can climb down in feature size, increasing the density of features and sell smaller chips with equivalent content.
“I can envison a chip at the bottom of a microtiter well, and between two wells, the entire genome — it’s compelling for high throughput analysis,” Rob Lipschutz, vice president of business development, told BioArray News.
Affymetrix, he said, is a company moving to its next phase of technology commercialization.
“There are a whole new collection of ideas that are going to come forward,” he said.
Already announced are new forays into DNA expression analysis and transcriptome research.
Protein arrays, however, are not.
“The technologies have to mature more,” said Lipschutz. “With DNA, you write down the sequence, design the ligand to detect that sequence — you can’t do that with proteins. I’m not saying that it won’t be done, it’s just technically challenging and it doesn’t lend itself to industrialization.”
In early 2003, Affymetrix will introduce barcodes into its product line, allowing users to use a handheld bar-code reader to track individual chips. The company is also licensing the APIs to its analysis data model.
David Craford, vice president market development, laid out a potential road map toward clinical diagnostics.
“In around five years, not just research MDs/PhDs but oncologists will get pressure from patient advocacy groups that these tools can help get more information. There is a growing body of data, especially in oncology, that opens the eye to what you can do using molecular technologies.”
“We built and validated the manufacturing facilities with [FDA] compliance in mind,” said Steve Karas, vice president and general manager of the company, and the factory chief.
The lesson learned?
The company has also learned from experience. In 2001, Affymetrix stock tumbled in March 2001 when the company reported that its U74 murine arrays contained defective sequence due to an error in reading the direction of sequence in the UniGene database.
“Take nothing for granted,” said Brad Kreger, manager of manufacturing process engineering at the plant. “We learned to do more rigorous checks of information.”
There too, is room to grow.
Affymetrix is spending an additional $1 million on R&D efforts in the fourth quarter. More growth.