At A Glance
Bob Searles, manager, spotted microarray core laboratory, and staff scientist, Division of Pathobiology and Immunology, Oregon National Primate Research Center.
Education: 1980 — BS, biochemistry, biophysics, Oregon State University.
1988 — PhD, biological chemistry, University of California, Los Angeles.
Chris Harrington, staff scientist, Gene Microarray Shared Resource; manager, Affymetrix Microarray Core, Oregon Health and Science University.
Education: BS — Biology, California Institute of Technology, Pasadena.
PhD — Molecular, cellular, and developmental biology, University of Colorado at Boulder.
Chris Harrington knows the Affymetrix microarray platform from three years of experience as a scientist and manager at the company and now another four years as manager of the Affymetrix microarray core facility of the Oregon Heath and Science University.
Industry insiders call Harrington a guru when it comes to quality assurance and quality control in processing GeneChips. And, at the recent annual meeting of the Association of Biomolecular Resource Facilities in Portland, Ore., Harrington’s lab presented a poster detailing the steps her lab takes to minimize assay variation and noise. These processes include the implementation of standard operating procedures recommended by Affymetrix — with modifications designed to limit variation and enhance sample throughput. Technical replicates (independent RNA labeling) show correlation values of greater than 0.97 in Harrington’s lab, even when the array assays are separated by six months or more.
Harrington is only a third of the team providing microarray-based analysis to OHSU researchers. The institution has two microarray facilities — an Affymetrix core, and the second, a spotted array facility — sharing wet lab space in the Vaccine and Gene Therapy Institute on OHSU’s West Campus in Beaverton, a suburb of Portland. The school also provides a bioinformatics and biostatistics core facility, which is located on its main campus. The three facilities are known collectively as the Gene Microarray Shared Resource and are directed by Klaus Frueh, an associate scientist at the Vaccine and Gene Therapy Institute.
The two microarray cores, situated together in a laboratory of approximately 600 square feet, provide an insight into the state of the portion of the microarray industry targeting the basic research market. In the decade since microarray technology was refined to the point where it could be consistently produced and distributed, countless hundreds of laboratories like this have been developed. The microarray market is dominated by the home-brew technology popularized by Pat Brown’s laboratory at Stanford University, while the Affymetrix platform dominates the smaller market for pre-printed arrays with an installed base of some 970 systems as of the end of the year. There is no reliable estimate of the number of laboratories that are using the home-brew technology, and there is no official census of the number of facilities utilizing these technologies in combination, such as OHSU does.
Whatever the numbers, the overall market is expanding but questions remain over when the basic-research segment of this market will reach saturation.
In the OHSU lab, on two benchtops sits an investment of over $200,000 in Affymetrix instruments, including a hybridization station, two fluidics stations, and a scanner, surrounded by brown bottles, glassware, Kimwipes, and gloves. One area, marked off by tape, is designated as a gloves-only RNAse-free zone.
In another area of the lab, a back room that the technicians call the “black hole,” sits a glass-encased robot and slide stackers. The room is darkened to better enable the gridding process for array images displayed on the computer screen just across from the spotter.
Here there is no market competition, just two laboratories using some creativity and following stringent protocols to provide their customers with data as precise as the technologies will allow.
BioArray News recently visited the microarray facilities of the laboratory, which fills one of many wet labs situated in a two-story brick building in an evergreen forest.
Harrington’s facility processed approximately 800 microarrays last year and is on pace to do 1,000 this year, as testified to by a jam-packed drawer in a benchtop where more than 400 GeneChips are stored after processing. The chips Harrington’s lab processes are mostly human and mouse arrays as well as yeast and rat arrays, with Xenopus and zebrafish chips soon to come, she said.
For processing GeneChips, Harrington goes by the book, Affymetrix’s GeneChip Expression Analysis Technical Manual, which tells “you what to do and where to buy the reagents,” Harrington said.
Harrington has a staff of four to assist her work. Kristina Vartanian is the lab manager, Brian Tompkins manages the data, and Hana Paik and Courtney Ryan perform the analysis as research assistants.
The Affy core facility is preparing to purchase Affymetrix’s new scanner, but Harrington said she is also planning on keeping the lab’s original scanner, the Agilent Technologies-manufactured model previously sold by Affymetrix. The company introduced its new scanner last year, requiring an upgrade — and a capital expenditure — for labs wanting to use its latest microarray products.
“[The new scanner] has been validated, and shown to produce good data,” said Harrington, “but for the time being, I’m keeping both — so I can be confident.”
She said some of the Affy core’s researchers are interested in using the company’s new one-chip whole-human-genome microarrays, requiring the lab to purchase the instrument.
For the Affy lab, Harrington said that labeling RNA is the rate-limiting factor in its processes.
“It takes two days to do,” she said.
“So much depends on the incoming sample,” she said. Some 15 percent of samples fail her lab’s first two quality control tests, she said.
The lab has created a script that extracts the “most salient metrics” that functions as another cross-check on experiments, one that has found a mistake made when one chip was scanned twice and filed under different names. The software flagged the results and the mistake was discovered, Harrington said.
The lab also utilizes freezers to store hybridization solution and save costs.
“Our hybridization is done in a cartridge-enclosed chamber,” said Harrington. “We don’t lose the solution, we draw it off and store it. For some investigators, we re-run it on other chips of a set or the newest [GeneChip] release. There is a limit, however, on how many times you can run it, and we don’t do it across the board.”
On the other side of the lab, the spotted array facility has adapted a hybridization technique where it prints on the middle of a slide and creates a sandwich of two chips, face to face, with the slight elevation of the bar-code labels providing just enough room for hybridization to occur.
“The variability goes down a lot, and it cuts down sample volumes by 75 percent, which saves money for the investigator,” said Bob Searles, manager of the spotted microarray core, which services the Oregon Cancer Institute as well as the university, and was created with a $700,000 grant from the Hedco Foundation of Northern California in 2000. Searles was conducting research on viral genomics and bioinformatics for the school’s primate center when the core microarray facility was organized and remains as a member of the center’s staff. Both managers are affiliated with the school’s Vaccine and Gene Therapy Institute.
Searles lab’s challenge, he said, is “printing a good slide, a nice round spot, with no donuts.”
The lab conducts its work on a five-day workweek with 8-hour shifts. In the three years the lab has been open, technician turnover has been low, with only one worker leaving — after being accepted for medical school, said Searles.
“The job can wear you down after a while; you have to do the same thing every day,” he said.
Searles has a staff of three — Julie Patch, David Lee, and Thaya Patton.
The spotted array laboratory provides services for 30 to 40 laboratories across the university and provides some 1,500 chips a year, printing arrays of approximately 8,400 genes twice per slide.
“Primarily, a handful [of investigators] use a lot of arrays on one project,” he said. “We think this year will be a breakthrough year in how many people use [the facility].”
Searles sees the spotted facility evolving away from printing its own chips, to purchasing pre-printed arrays.
“Printing is a siphon of time,” he said.
The lab wants an affordable chip that could contain viral and host genes on a single slide, which is currently provided by the core.
“We don’t see that yet in the market,” Searles said.
The spotted lab pegs its chip costs at $60 each. “The costs are paid by the individual researchers with labor costs provided by grants,” he said.
Print runs take 27 hours to conduct on the lab’s three-year-old “much modified” Cartesian robot, which is under a service contract costing $25,000 a year. Service calls are conducted with Searles, phone cradled between cheek and shoulder, talking to the company.
The lab uses only Corning UltraGaps slides for its printing.
“They are the only ones we are happy with,” Searle said. “They last a long time.”
The lab uses TeleChem pins for its printing and expects between a year to two years of service from each pin. Searle said the lab is also considering purchasing pins from Point Technologies for cost savings.
The lab has started to design a laboratory information management system on the Thermo’s Nautilus platform, which Searle and Harrington said has them looking forward to providing more efficiency in tracking the laboratory’s activities.
Searle has also collaborated with the bioinformatics core facility to create a quality control module that targets spurious spots that arise in hybridization. When implemented, the system identified occasional defects that led to the lab creating an extended wash cycle for its arraying pins.
“Chris and I serve an almost non-overlapping set of users, so we don’t really compete,” Searle said. “We do find that we learn a lot from each other, thereby improving the quality of the results that each core produces.”