With the ever-increasing density of microarrays comes a need for moresensitive and high-throughput ways to read the resulting data. Arrays contain tens of thousands of 50- to 150-μm-diameter spots per square centimeter, and the vast amount of information they produce has pushed imaging and scanning equipment to its limits.
Several recent advances promise to improve the sensitivity, speed, and accuracy of microarray readers. Scanners are becoming more versatile, able to read a broader range of chips, substrates, and detection molecules, and they are evolving to meet the demands of increasingly high-throughput technologies.
Resolution and sensitivity are the defining characteristics of microarray readers. In general, resolutions of 5–10 μm are needed to detect transcripts with a relative abundance of one copy in 500,000, or three to 10 copies per cell. And the presence of at least 2–5 molecules/μm2 is necessary for low-level fluorescence detection.
"The current trend in...
"The ability for the scanner to read any 2.5 x 7.6 cm chip really relates to the trend of cross-platform design happening within the industry," says Saunders. "People can use an Agilent array, arrays from any other vendor, or homebrew arrays in the scanner; they're not locked into a closed system," he says. (That's in contrast to systems from such vendors as Affymetrix of Santa Clara, Calif., and Applied Biosystems of Foster City, Calif., whose readers accept only their own chips.) "In addition, upgrades to the scanner's new feature-extraction software means that homebrew array users no longer have to feature-extract using another package," he adds.
GE Healthcare (formerly Amersham Biosciences) of Piscataway, NJ, has also adopted the goal of versatility. "Microarray technology has gone beyond the research stage of development and has crossed the chasm into the mainstream market to the extent that we are putting microarray capabilities onto our general-purpose laboratory scanners," says Brian Ayling, director and product manager of gene expression at GE Healthcare.
The company's Typhoon 9410 Variable Mode Imager became the first large-format gel and blot imager that could also image microarrays. "People are familiar with using these kinds of instruments for imaging gels and blots. This means that generalists don't have to make a significant investment in specialized microarray equipment," says Ayling.
Scanners must also meet the demands of increasingly high-throughput technologies, such as Affymetrix's new 96-well HighThroughputArray (HTA). Axon Instruments, based in Union City, Calif., recently introduced the GenePix Autoloader 4200AL, an add-on to the company's GenePix Professional 4200A microarray scanner. The autoloader loads up to 36 standard microarray slides to automate the processes of scanning and analysis. Axon will also supply scanners to Affymetrix for the company's 96-well HTA technology.
Likewise, Alpha Innotech, San Leandro, Calif., recently released its white-light AlphaArray 8000, which uses CCD-detection technology coupled with a broadband excitation source and compatible with a wide variety of fluorophores. The technology provides resolution images down to four microns and accommodates standard slides as well as custom formats and microplate based arrays.
Tecan Group, based in Maennedorf, Switzerland, has developed what the company calls a "highly flexible scanning technology." It allows researchers to scan many different samples automatically, from glass slides to microtiter plates, on the same equipment without having to switch scanners or make extensive adjustments each time the format type or size changes. According to information on the company's Web site, the new LS Series of laser scanners can operate up to four excitation lasers and 28 emission filters.
"Most of the major hurdles in microarray scanning for DNA have now been overcome, and there are a range of good systems and analysis tools on the market," notes Richards. "The introduction of reliable and affordable three-laser systems such as the Genetix aQuire have allowed the use of a greater range of dyes such as fluorescein and green fluorescent protein," she adds.
Samples analyzed by microarrays are generally fluorescently labeled, often using probe families with similar chemical structures but different spectroscopic properties, such as the cyanine (Cy) dyes Cy3 and Cy5. But fluorophores present problems, including photobleaching and spectral overlap (or bleed-through), which can diminish the quality of collected data.
Because of these problems, some foresee a move away from fluorescence detection. "I think we are going to see more alternate-detection strategies, which may result from developments in the protein array industry," says Todd Martinsky, executive vice president of TeleChem International, Sunnyvale, Calif. "Some of these systems might not be better technically, but they may be adequate for a particular application. For example, colorimetric detection certainly isn't as quantitative, but not everybody needs quantitation," he says.
TeleChem's Spot-Ware Colorimetric Microarray Scanner allows researchers to use horseradish peroxidase, alkaline phosphatase, gold-silver developers, secondary antibodies, and any other labeling system that produces a pigmented reaction precipitate. According to Martin-sky this scanner will read DNA and protein, and costs about $4,000. "Some people just need a yes or no answer and don't need to spend $40,000; you are tailoring a particular detection strategy, and that reduces costs," he says.
Applied Biosystems' new Microarray Expression Analysis System combines chemiluminescence and fluorescence as a dual method for detecting gene expression in biological samples. Each spot on the arrays contain a 60-mer and a 24-mer oligonucleotide. The former is labeled with digoxigenin for chemiluminescent detection, the latter with a fluorescent dye. In the absence of a chemiluminescent signal, or the presence of a very low signal, each of the approximately 35,000 spots can be precisely located through the fluorescence signal. This allows the system to find each spot and reproducibly map it, and if there is a lack of correlation, the spot is rejected. The Expression Analysis System also includes a CCD camera and high-powered light-emitting diode for its "pseudo-fluorescent" illumination system.
Resonance Light Scattering (RLS) is another relatively new detection strategy for microarrays. The ultrasensitive signal generation and detection technology, developed by Genicon Sciences, San Diego, (now owned by Invitrogen, Carlsbad, Calif.) can be applied to a wide variety of molecular assay formats and platforms. The company's GeniconRLS System can generate, detect, and image one- or two-color gene-expression microarrays using as little as one to two micrograms total RNA for 10-fold increased sensitivity over fluorescence. In addition, it can detect twofold differences in gene expression and perform multiple readings without signal decay, the company says. Qiagen, based in the Netherlands, has released its HiLight Array Detection System based on the technology.
THE NEXT WAVE
Researchers at Sandia National Laboratories and the University of New Mexico, Albuquerque, have developed an alternative strategy, called hyper-spectral scanning, which could soon surpass current fluorescence systems. The scanner achieves resolutions between 3 and 30 μm and records the emission spectrum between 490 and 900 nm with a spectral resolution of 3 nm for each microarray pixel. According to a recent report, this spectral information, "when coupled with multivariate data-analysis techniques, allows for identification and elimination of unwanted artifacts and greatly improves the accuracy of micro-array experiments."1
David Haaland, who pioneered the hyperspectral scanner at Sandia, notes that it quantifies every emitting source, including those not associated with the labels, such as those from the buffers used in printing the DNA microarrays. "It separates them out at every pixel in the image. Then when coupled with multivariate curve resolution, the relative fluorescence of all fluorescing components can be quantified to separate out the effects of the contaminant and glass emissions from those of the fluorophore, resulting in a more accurate image of the DNA labels," he says.
Haaland says the microarray field holds great promise, "but I think some of the experimental variance in the data makes it difficult to see the biology. These new detection approaches make the whole industry better and will allow scientists to get more information."
Selected Suppliers (63K)