1-888-WINSHIP
A microarray, in the simplest terms, is a regular series of microscopic probes arrayed on a membrane. The membrane (frequently a glass slide) is then treated in some manner, and the individual probes are assayed to determine their response to the treatment. This technique has been used to stain multiple tissue samples at once, or to determine how an antibody reacts to multiple protein types. But the most common type of microarray, and the type that we use at WCIMC, is the DNA microarray [1]
A DNA microarray is typically used to detect which genes are expressed in a sample. Total RNA is extracted from the sample tissue. Message RNA is primed with anchored oligo-dT, reverse transcribed, and cRNA is labeled with a tagged molecule. This labeled cRNA is then allowed to bind to its complement on the microarray, and the amount of bound target is detected by measuring the chemiluminescence or flourescence. If the probe has been properly designed and the hybridization is stringent enough, the only target that will bind to it is the message for a specific gene, and so we can use the relative signal of the probe to infer gene expression.
It is tempting to consider microarray results as quantitative - after all, if the measured signal is 6 times higher in the experiment than the control, that must mean that 6 times as much cDNA bound to the probe, and so the message for that gene is 6 times more abundant, right? Unfortunately, there are a number of sources of potential error inherent with microarray technology. Because of these errors, we regard all microarray data as an indicator of the underlying processes, and strongly recommend that users always verify any important microarray results with an alternative technology, such as Real-time PCR. Most journal reviewers insist on seeing this verification performed.
As microarrays have become prolific, a number of vendors have introduced their own line of arrays. These different platforms, while grounded in the same basic principles, differ in their exact implementation. Despite these differences, they all work basically the same way, and the data is analyzed the same way - unless the number of samples per array changes. In order to get a more direct comparison between two samples, without being affected by slide-to-slide variation, two-color microarrays have traditionally hybridized two samples to a single array. Each sample is labeled with a different dye, and the signals are detected at different wavelengths. While this method does make direct comparisons simpler, it does have drawbacks. Signals can be skewed from competitive hybridization, comparing more than two groups becomes complicated, and every desired comparison requires more RNA to be isolated and another array to be hybridized. Because of these drawbacks, most commercial vendors design their arrays to be used with only one sample. Array-to-array variation is minimized through careful manufacturing and data normalization, and all comparisons are performed with a computer, after each sample has been hybridized. Single-color can thus be a more expensive, if you have to buy more arrays. However, two-color arrays often use a universal human reference RNA in the control channel and can require dye-swapping experiments, which can result in more arrays for a two-color experiment than a single-color experiment. Because the type of platform you choose influences your experiment design, we recommend choosing a platform as quickly as possible.