Microarray experiments are extremely powerful and provide researchers with a new and exciting means of tackling important problems on a genomewide scale. Most microarrays contain probes for 10,000–40,000 different genes, allowing researchers to assess simultaneously changes in expression of nearly all the genes in the genome. However, they are also complex, time-consuming, and often very expensive experiments, and they generate large and complicated data sets that require substantial effort to analyze and validate. For these reasons, researchers should not be lured into performing microarray experiments without spending some time considering other options or without considerable thought regarding appropriate experimental design. New users should
consult extensively with their local microarray core facility before beginning to prepare samples for microarray analysis. Every microarray facility can tell stories about users who approached them with samples only to find out that unsuitable preparation or storage had resulted in RNA that was too degraded for high-quality analysis. Proper preparation and storage of the RNA is crucial to the success of microarray experiments.

This is especially true for samples derived from patients or tissues that are difficult or impossible to replace. The microarray facility should be able to guide users to the best
methods for preparing samples and storing the RNA to ensure that their experiments will succeed. Because of these limitations, some experiments are better suited for microarray analysis than others.

Microarray technology has proven to be extremely powerful for following changes in gene expression that occur as synchronized cells progress through the cell cycle or when tissue culture cells are treated with a drug or are infected with a virus expressing a recombinant transcription factor. In such situations, all the cells in the population are responding in parallel and relatively synchronously, and the microarrays, which measure the average change in gene expression in the population of cells being studied, can detect changes in gene expression that occur simultaneously in all the cells. Because of variations in measurements, microarrays are best at detecting changes that are relatively robust—a twofold or greater change is a common benchmark—in genes that are expressed at relatively high levels. Cells from different individuals, such as different patients, can display markedly different gene expression patterns, so microarrays perform best when the samples are closely related, such as tissue culture cells or treated vs untreated cells from a single patient or animal. Because different cell types display complex differences in gene expression patterns, heterogeneous samples, such as solid tumors or tissue samples, give complex microarray results. Optimum results are obtained from homogeneous samples, such as cell lines or purified cell populations, when they are available.