| A Bio-IT World special section
Biochips for Gene Expression
Microarray manufacturing technologies are giving new meaning to the term 'custom.'
By Mel Kronick
October 10, 2003 |
DNA microarray technology has revolutionized the study of gene expression and drug discovery, but the upfront investments traditionally required to create new microarrays are increasingly at odds with the realities of modern biology. As more sequencing studies are done, the annotation of sequences is changing rapidly. DNA sequences believed yesterday to represent one protein could today represent two or more; sequences thought to correspond to several proteins could be alternative splice forms of the same gene.
The limitations of technologies historically employed to manufacture microarrays have made it impractical to update microarrays based on new knowledge. The two leading processes for microarray creation today are still expensive and time-consuming tasks: photolithographic techniques require the design and fabrication of expensive masks; deposition of presynthesized oligos or cDNA-derived polymerase chain reaction (PCR) products demands the creation of oligo or clone libraries.
Alternatives have become both necessary and available, and their invention is paving the way for new microarray applications and redefining the concept of a "custom" microarray.
|Just as an inkjet printer can print 20 copies of one page at the same cost as printing 20 unique pages, new microarrays can be profitably created without bulk orders or large set-up expenses.
Building on the early work of Ed Southern and colleagues at the University of Oxford, new approaches have been developed to enable the on-demand in situ creation of oligo microarrays. The technologies that enable this capability include inkjets that spit out individual nucleic acid amidite monomers, micromirror devices using light to deprotect monomers without masks, and novel electrochemical synthesis methods. These methods obviate the need to prepare any physical, biological, or chemical entities beforehand. Just as an inkjet printer can print 20 copies of one page at the same cost as printing 20 unique pages, new microarrays can be profitably created without bulk orders or large set-up expenses.
This revolution in technology is just now beginning to take hold, providing scientists with lower-cost custom microarrays. Researchers can rapidly and affordably iterate their designs and take full advantage of the "best view now" with the most up-to-date genome sequence knowledge. As the boundary of inflexible content is removed, new research methods are becoming more broadly available.
For example, Dan Shoemaker and co-workers at Rosetta Inpharmatics pioneered one of the earliest applications of maskless in situ arrays. Rather than predefining genes of interest in their microarray design, probes were tiled along the entire human genome.
Only those sequences that were actually expressed in the tissues of interest fluoresced, such that the true biological definition of what was a gene could be directly observed, not just inferred, and the annotation could be updated immediately. To examine one particular region in more detail, a higher resolution tiling of that region was quickly produced that clearly showed molecular-level details such as exon-intron borders.
Similarly, new technologies better enable the iterative study of splicing functions, where a scientist wants to design an explicit variety of probes to understand the presence of splicing or isoform expression in specific genes. Inflexible array content would limit the questions that could be asked. With on-demand in situ microarrays, specific research needs can be more easily accommodated.
Custom Becomes Standard Issue
As more genomes are sequenced, researchers want to apply microarrays to the study of plant genomes, model organisms, and microorganisms. While once this might have required a custom order, maskless in situ microarray printing permits the economical creation of commercial microarrays for particular organisms that might be of interest to only a small number of researchers. In bypassing the need to clone out cDNAs and amplify PCR products, oligo microarrays have been developed and are being used for such idiosyncratic, albeit important, organisms such as the rice fungus, Magnaporthe grisea.
The affordability and flexibility of new manufacturing techniques also improve the quality of microarrays by enabling iterative design and wet-lab testing prior to mass manufacture. Think of Jeffrey Waring and colleagues at Abbott Laboratories creating a library of genes they knew were being highly expressed in rat liver tissue under hepatotoxic conditions.
After computationally selecting good probe sequences for genes of interest, the probes were empirically evaluated on a small set of microarrays to improve annotation, eliminate or reduce cross-hybridization, and ensure robust detection performance. Then, the updated design incorporating the best probes was produced in the larger quantities needed for experimentation. This same iterative process can be applied to the manufacture and the updating of commercial microarrays.
When old designs are needed to repeat an experiment, there's no need to keep extra inventories around; the layout and sequence files for each microarray design need only be stored. This type of economy will revolutionize the large-scale production of arrays. It will soon become as easy to order microarrays as it is to order oligos. Already some companies are quoting five-day turnaround time once the design has been submitted.
Next Generation: One Slide, Eight Experiments
As the technology matures, further possibilities of the maskless methods have become apparent. While high-feature-count microarrays are ideal for screening entire genomes, there are many applications of increasing interest — e.g., patient stratification, toxicogenomics, metabolic studies — where a few thousand genes or less would suffice.
For instance, Agilent Technologies researchers at the 2003 annual meeting of the Association of Biomolecular Resource Facilities (ABRF) presented a standard 1 x 3-inch slide segmented into smaller arrays. Just as a printer can fill a page with letters or print discrete paragraphs, this slide included eight identical arrays. Onto each small array a different sample was placed, allowing multiple experiments to be run on one slide. Using this approach, 200 patient samples could be analyzed on 25 slides, and less labeled material would be needed for each array. The cost per experiment for such studies is thus considerably reduced.
In another example of time, cost, and application improvements, Paul Meltzer at the National Human Genome Research Institute (NHGRI) and colleagues are creating design methodologies, assay protocols, and analysis methods to quantitatively characterize instabilities in genomic DNA accompanying tumorigenesis.
Probes might be located at approximately 50-kilobase intervals to discover regions throughout the entire genome where changes in DNA occur during cancer progression. When deleted regions are observed, the investigators could then "zoom in" by synthesizing new arrays with a nominal resolution as high as 60 base pairs or less and define the region in more detail. The flexibility to tailor the design to each problem like this was unimaginable with microarrays of the past.
Today's applications are no longer limited by what is commercially available or by what previously took many weeks and many thousands of dollars to create. Increasingly, microarray applications are limited only by the imagination of the researcher.
Mel Kronick is chief scientist of the BioResearch Solutions unit within the Life Science and Chemical Analysis business of Agilent Technologies, in Palo Alto, Calif. He may be reached at email@example.com.