Methods for the analysis of DNA have long been recognized as ligand assays, relying on the binding of target molecules to a specific recognition reagent. Techniques for immobilizing DNA on nitrocellulose paper and for detecting the fixed nucleic acid with radioactive probes were widely adopted.9,141,142 Building on the idea of using immobilized probe collections, as in the ''reverse dot blot,'' Southern developed a method for the parallel, in situ synthesis of oligonucleo-tides using standard nucleotide synthetic reactions as a way to generate oligo-nucleotide probe arrays on microscope slides for highly parallel hybridization analysis.20,143,144
DNA microarrays are assemblies of nucleotide sequences, each of which represents a single gene, splice variant, or another DNA element, attached to a solid surface. Many variations on this central theme have appeared since the origin of the concept. For example, DNA microarrays are distinguished from earlier hybridization methodologies such as reverse dot blots by their relatively high probe density and by their use of miniaturized, solid, nonporous supports. Rigid supports, particularly those that are optically transparent and thermally conductive, are more practical than flexible supports. They are also more practical for interfacing with automatic fluid delivery or printing equipment, and they best accommodate the automatic scanning systems that are used to image arrays. The method of generating probe sets has also been modified in several different ways. In one successful modification, photochemistry and photolithography were incorporated into oligonucleotide synthesis,21 and in another, piezoelectric nozzles and ink jet heads "printed" DNA synthesis reagents directly onto substrates for in situ oligonucleotide synthesis localized by surface chemistry.145 Thus the idea of generating microarrays, or "gene chips," capable of high-throughput screening took root.
As the amount of available information about the sequence of the entire human genome has emerged, the density of successive generations of DNA microarrays has continued to climb. Recently, several companies have compressed the entire protein-coding portion of the human genome onto a single array—that is, some 30,000 gene sequences are represented on one slide. The result is a dramatic increase in information gathered from each profiling experiment affording the possibility of greater insight into complexities of cellular biology.
Measurements made using DNA microarrays are difficult to optimize to a tightly specified standard because of the high density of information and its parallel nature. It is equally difficult to validate individual probes globally in mi-croarray hybridization assays at every data point, since it is impractical to check every point by a reference assay method. As a result, doubt has persisted about the value of the massive amount of data that microarrays produce. Recently, however, a trio of studies represents a systematic attempt to assess the reliability and reproducibility of microarray data across platforms and between laboratories.146-148 In each case, by using standard protocols the picture presented is shown to be better than previously thought. They found, first, that both commercial and homemade arrays could deliver good results in experienced hands. They also found that standardized protocols largely solved the problem of poor reproducibility, and finally they found that the majority of microarray data reflects the underlying biology, even when different platforms are used to assess gene expression. On the other hand, the picture was not perfect and these studies called attention to some problem areas or issues. For one, they note that if the investigator is inexperienced, or does not have access to a core facility, use of a commercial chip should be considered. Second, if collaboration with another investigator is possible, the investigator should ensure wherever possible that the same platform and a common set of procedures are used. And third, if comparisons across platforms, say for gene expression, are necessary, a more biologically meaningful and statistically robust approach is to make relative rather than absolute comparisons.
The first use of microarrays was in immunological assays for diagnostic pur-poses,149 but the growth of genome sequencing created a demand for high-throughput, parallel multigene analysis. The uses of DNA microarrays for nucleic acid analysis have come to occupy niches in nearly every area of basic biological research and medical science. Direct sequence analysis of SNPs and entire genomes were made that incorporated both comparative and quantitative measurements. The design and fabrication of microarrays for applications as diverse as genome mapping,150 genotyping and reference-based sequence checking (rese-
quencing), , and gene expression profiling were also among some earlier applications. The technology continues to develop and is being transferred into other areas as in the development of proteomic, glycomic, and tissue arrays152 and G-protein-coupled microarrays,153 all of which carry the potential as aids for drug discovery. In the biomedical field, potential applications of this technology include the assessment of RNA and protein alterations as diagnostic markers in the clinical arena, particularly in oncology.154
In 1999, an appreciation of the different types of microarrays was chronicled in the first of the special series entitled The Chipping Forecast.74 In 2002, the second in this series focused on new applications of microarrays.75 A third part in this series published in 2005 dealt with perspectives on the application of microarrays to the assessment of SNP variation, large-scale structural variation, identification of epigenetic markers, bioinformatics of microarray data, and RNA interfer-ence.155 A few current examples of applications of DNA microarrays of particular interest to pharmacogenomics are considered below (pp. 214 and 215).
Was this article helpful?