Practical Examples in Pharmacogenomic Analysis

Each of the tools and databases discussed previously is grounded in a significant amount of both technical and theoretical detail. To illustrate the utility of these tools, practical data-analysis examples are provided that outline how a microarray experiment can be designed and analyzed. In addition, the annotation of an uncharacterized EST is defined and mapped to the genome. These examples also demonstrate how to assign annotation to microarray analysis, when the identify of a EST represented on an array may be unknown. 

The most robust method for identifying alterations in transcript levels is the WRST. This method does not make assumptions concerning the distribution or the variance of the data to be compared. However, many investigators use the /-test due to its simplicity. Alternatively, a model-based /-test from the GLMM could be used which performs contrast tests between two groups of interest, such as an effect at a particular dose vs. a vehicle effect, much like a /-test. The values for the test come from the model fit to the data, and provide a better estimate of the difference between the two groups, as other data are used to better constrain the real effect of the dose and vehicle under consideration. 

When constructing the GLMM to perform the model-based /-test, the terms within the model include the dye (to model dye-effects), the microarray (to model any random microarray-specific effects), and the dose (including vehicle, to model dose/vehicle effects). 

Does This Sequence Match Any Known Mouse Genes  

Where Does This Sequence Match to the Genome  

---