In indirect experiments, samples are hybridized to a common reference RNA, and then the two separate experiments are compared to each other (A/ref versus B/ref). The reference RNA is typically the same preparation of mixed stage hermaphrodite polyA+ RNA.
Use indirect experiments when each experiment is compared to many other experiments. For example, to compare many mutants to each other and to wild-type, one could compare mutantA/ref, mutantB/ref, mutantC/ref etc. versus WT/ref. Another example of an indirect experiment is to compare a mutant time course (or dose response) to a wild-type time course (A1/ref, A2/ref, A3/ref ... versus WT1/ref, WT2/ref, WT3/ref ...).
The advantage of an indirect experiment (A/ref versus B/ref) over a direct experiment (A versus B) is that each sample is done once using an indirect experiment. Using a direct experiment, a repeat experiment would need to be done to compare each sample to a new sample. The disadvantage of an indirect experiment is that there are two hybridizations instead of one for each comparison, so that the standard error for iindirect experiments is greater than that for direct experiments.
To determine which genes are significantly different between two points in an indirect experiment, use a Student's t-test for two populations.
To determine which genes change within a series (such as a time course), use a one-way analysis of variance (ANOVA).