Microarray experiments reference sample

Another important point was made by this study if possible, use reference samples with gene expression levels similar to those expected for the test samples. This should improve quantification at the low end. Moreover, the Tan et al., (2003) and Barczak (2003) studies also demonstrate the need for the standardization of commercial microarray products and reference standards so that both interlaboratory and cross-platform performances can be properly assessed. While commercial microarrays from different vendors all claim substantial gene probe "real estate," the number of probes in common is smaller. This makes it difficult to compare experiments across platforms, especially those from major suppliers such as Affymetrix, Agilent, and Amersham (Figure 2.8). 

When two dyes are used, there are two experimental approaches (Figure 3.2). In the most common approach ( reference design, Figure 3.2), the experimental sample is labeled with one dye while the common reference sample is labeled with the second dye, and both are hybridized to the same microarray. Alternatively, for a more statistically balanced and efficient experiment, samples can be hybridized with other samples in a loop design (Figure 3.2). In the loop design, the two experimental samples (no common reference) are labeled with the two different dyes and hybridized to the microarray. 

The use of a common reference RNA allows for the comparison of data between various array experiments. Typically a pool of RNA derived from a variety of tissues is used as a reference sample. Efforts are being made to implement common standards (MIAME, minimal information about a microarray experiment) for transcript profiling through the MGED Society to enable data sharing between different groups [50]. 

---