×

Post-processing multi-treatment time-course microarray studies using the Johnson-Neyman procedure. (English) Zbl 1228.62145

Summary: The decline of microarray prices has motivated more complicated multi-treatment time-course experimental designs with multiple replicates at each time point. Depending on the design, researchers may be interested in both differential expression with respect to time and treatment, as well as their interaction effects. Careful examination of several regression based methods for the analysis of multi-treatment time-course experiments reveals that the current implementations can be expanded to facilitate the interpretation of results. We address the postprocessing of regression analysis results and, in particular, focus on interactions between treatment and time, where effects are not additive. For this purpose, we directly examine differences between regression curves of two treatment levels across time using the P. O. Johnson and J. Neyman [Statist. Research Mem. 1, 57–93 (1936; JFM 62.0633.08)] procedure on sample data sets. This approach is useful for organizing a complex body of results and can help investigators answer questions of interest.

MSC:

62P10 Applications of statistics to biology and medical sciences; meta analysis
92C40 Biochemistry, molecular biology
62J05 Linear regression; mixed models

Citations:

JFM 62.0633.08
PDFBibTeX XMLCite