Figure 1. Preparing to use real-time PCR array technology, Emi Arikawa, PhD, a staff scientist at SuperArray, loads an array to analyze pathway gene-expression. (Image courtesy of SuperArray.)

Roche NimbleGen's target-enrichment advance appeared in the November 2007, Nature Methods. This approach, though, remains in development. "The basic idea is that an array is designed with probes that recognize specific regions of interest in a genome," says Rose. "The genomic DNA is then hybridized to the array, the non-targeted regions are washed away, and the targeted regions are eluted in a form that can directly feed into a shotgun-sequencing pipeline." As a result, scientists can collect specific regions of DNA. In addition, this target-enriching technique lets researchers sequence contiguous areas of the genome suspected as causes of disease.

In general, the recent and rapid advances in microarrays arose from a combination of simultaneous events. "It's the perfect storm as far as technology goes," says Jessica Tonani, MS, genotyping specialist at Affymetrix in Santa Clara, CA. "The technology is affordable and mainstream." As a result more scientists can use microarrays, which leads to the discovery of ever-more applications.

Current advances show that the benefits of today's microarrays go beyond simply bigger arrays. "What is actually happening," says Xiao Zeng, PhD, senior director of R&D at SuperArray in Frederick, MD, "is that scientists are aiming microarrays at a deeper level of questions in terms of biology." So microarrays do not just capture more information, they can now gather information that also contains higher content. Moreover, scientists know that results from one microarray experiment often lead to others. For example, Illumina in San Diego, California, recently developed its BovineSNP50 BeadChip. Researchers used Illumina's Genome Analyzer to identify specific areas along the bovine genome that they wanted to target for additional fine-mapping studies using a customized array developed by Illumina. So one microarray experiment can allow more-specific explorations.

Do it with more density

Figure 2. NimbleChip HD2 microarrays include 2.1 million probes for experiments in genome-wide detection. (Image courtesy of Roche NimbleGen.)

Today's microarrays also add power to studies, because they give more coverage of variants, and it's more affordable to run large numbers of samples. "Studies used to be underpowered," Tonani says, "because scientists could not run enough samples or could only look at about half of the potential variation." Now, scientists can follow as much as 90% of the potential variation and run thousands of samples. Some Affymetrix expression microarrays also let researchers look at exons, instead of just looking at the ends of genes. Depending upon the exons spliced together, a gene can produce different products. "If you look at drug metabolism, that could influence how you metabolize a compound," says Tonani, "but if you were just looking at the ends of a transcript, one variant might look the same as another."

Exploring variations in copy number also teaches us more about the structure of genes. "Looking at copy-number variation gives one the power to identify risk factors for specific diseases," says Jay Kaufman, senior director genomics marketing at Agilent Technologies in Santa Clara, CA. "It also allows people to correlate clinically relevant copy-number variations with phenotypic information to identify underlying causes of genetic-based disorders."

Many scientists study copy-number variations. The Wellcome Trust Sanger Institute runs a project on copy-number variation, and the website for this project (www.sanger.ac.uk/humgen/cnv/) states: "This type of mutation has often been overlooked in previous surveys of mutations that cause genetic diseases. We do not know what proportion of genetic disease is caused by copy number variation, but we suspect that it is appreciable."

Illumina also recognizes the importance of copy number-variation content. According to Carsten Rosenow, PhD, senior marketing manager of DNA analysis products at Illumina, the company's new Infinium HD Human1M-Duo and Human610-Quad BeadChips both include high-value, proprietary copy number-variation content developed in conjunction with deCODE genetics (Reykjavik, Iceland). Moreover, both of the new chips pack 2.4 million probes. These new arrays take on multiple samples simultaneously two for the 1M-Duo and four with the 610-Quad, as the names imply. These microarrays require less sample that previous chips, and run faster, says Rosenow.

Dealing with the data

Figure 3. With the Affymetrix Genome-Wide Human SNP Array 6.0, researchers can explore more than 1.8 million markers of genetic variation, including single nucleotide polymorphisms (SNPs) and copy-number variation. (Image courtesy of Affymetrix.)

In fact, some of the information gathered in the past even includes errors that scientists must address. "Although the human-genome sequence has been worked out for quite a while," says Zeng, "lots of the annotation was not quite correct." He adds, "There are 50% false positives in some cases. That's kind of alarming, especially when you don't know the biological information that you are getting."

Part of the trouble is that microarrays from different suppliers do not always produce data that correlate well. That becomes problematic when scientists want to do different types of microarray experiments to look at the same problem, say the cause of a specific form of cancer. "Does the platform that you have enable you to do that?" asks Kaufman. "If not, what do you do?"

One thing researchers could use is a standardized set of analysis tools for microarray experiments. "There are various tools and algorithms that people use," Kaufman says," but standardizing them might not be easy to accomplish until there are mandated guidelines for that."

Scientists need those tools to handle archived and new data. So standardized algorithms must handle data collected in old and new ways. Likewise, scientists often want to share data. But, as Kaufman says, "There are many challenges in data sharing and database mining."

Fueling flexibility

Figure 4. SuperArray's RT² Profiler PCR Arrays accurately profile gene expression for a biological pathway or disease using real-time PCR. This shows the amplification curves from a PCR Array representing a panel of 96 genes from a biological pathway. Over 100 RT² Profiler PCR Arrays are available for expression profiling in immunology, cancer, signal transduction, and other applications. (Image courtesy of SuperArray.)

To give scientists even more flexibility in microarray experiments, Agilent developed its eArray interface. "This is a web-based tool that allows researchers to select from a large database of existing microarray probes, make use of their own custom probes, or use a combination of both," says Kaufman. This way, scientists can simply go online to design their own microarrays. Agilent released version 5.0 in November 2007, and this company has plans for additional releases throughout 2008. "Our database will soon surpass 12 million probes," says Kaufman.

Illumina's iSelect Infinium Custom Genotyping also adds flexibility to microarray experiments. With this 12-sample BeadChip format, a customer selects the desired SNPs for a microarray. "One customer wanted to screen for cardiovascular disease," says Rosenow, "and they gave us 60,000 SNPs and asked for an array that included just these." Illumina also aimed the iSelect technology at agricultural applications. In January, the company released its Infinium BovineSNP50 BeadChip. "This microarray can identify specific loci in cattle, such as markers for marbling of the meat." says Debora Bailey, MS, Illumina's product manager DNA analysis products.

Scientists also keep putting new types of probes on microarrays. microRNA (miRNA) keeps attracting more attention. These short, non-coding sequences are known to be involved in regulating gene expression. As a result, researchers want better ways to track the activity of miRNA. In April 2007, Agilent released its first miRNA microarrays, and it is developing a second-generation product, says Kaufman. "It will be an update based on the expanded human content in the Wellcome Trust Sanger Institute's miRBase, as well as adding mouse and rat miRNA array products."

More just ahead

The combination of SNP and copy-number probes on the same microarray also points to the future. Most of today's arrays look at one form of probe. There are microarrays for DNA, and ones for protein expression. Other arrays look for gene methylation. Promoter arrays study regulation. To study gene knockdown, scientists turn to small-interfering RNA, or siRNA, microarrays. And the list goes on. "We see the microarray philosophy expanded to include some way to handle all of these needs simultaneously instead of in series," says Martz. "This will expand platforms to new applications."

The future of microarrays will also bring many new applications. "There's a movement toward using microarrays in diagnostics, especially related to the cytogenetic area," says Kaufman. Microarrays could one day be used to screen fetal DNA for gene-based diseases or developmental disorders. Arrays eventually may be used to determine the specific drug that would work best for a disease in people with a specific genotype. As these advances develop, we will see microarrays evolve into biological minicomputers of sorts a long way from the simple switch detectors they once were.