|
The following examples demonstrate how to carry out wildtype vs. mutant experiments and time-course experiments, supposing you have 8 cDNA microarray slides
Scenario 1: Wildtype vs. mutant experiments
If you want to compare gene expression profiles of a wildtype strain (R) and two mutant
strains (M1 and M2), one way to do that is to compare the wild type with each mutant,
respectively, as shown in figure a. Each comparison can have four biological replications
in order to maximize the resulting inferential power.
If, in addition to comparisons between wild type and mutant, you are also interested in the
comparison between the two mutant strains (i.e. the mutans are related in some manner), then you
should use the design shown in figure b, which is a circuit design. In this desgin, each pair of
samples are compared directly with each other, with two biological replications. Six slides are
used in figure b. You could then use the remaining two slides for additional comparisons of the
pairs you are most interested in, or use them as a backup in case some hybridizations are unsuccessful.
Scenario 2: Time-course experiments
In time-course experiments, the design choices depend on the comparisons of interest.
Design c in the left uses T0 as a common reference, whereas design d involves hybridizations
between consecutive time points as well as hybridizations between T3 and T0.When the main focus of the experiment is on
the relative changes between T1, T2, T3 and the initial time point T0, design c is the better
choice. However, if more subtle variations from one time point to another are of greater
interest, then design II will be preferable. This is an illustration of how the comparisons
of greatest interest determine the best design.
|
