投影片 1

Reference design and Balanced block design

Balanced block design (efficient)

Balanced block design can only be used only for simple class comparison type of analysis (looking for genes that are upregulated or downregulated between two samples).

Reference design (flexible)

Reference design is far superior for more complex class comparison as well as for class prediction or class discovery where clustering analysis is required. In fact, clustering analysis can not be performed on experiment conducted using balanced block design. As a rule reference design should be used when amount of material is a limiting factor, and balanced block design should be used when number of slides is limited.

Loop design (elegant)

Reference RNA for reference design experiments used does not necessarily need to be physiologically relevant to your RNA. However, reference design should have at least 80% coverage for the particular slide (i.e. should contain mRNA representing at least 80% of genes on the slide). RNA isolated from a cells or tissues usually has 50-60% coverage. Reference RNA is, most often, a pool of RNA from several different cell lines or tissues. Human, mouse, and rat universal reference RNAs are available from several vendors .