Power and sample size determination for time course microarray differential expression studies: a positive false discovery rate and permutation-based simulation method

Microarray experiments enable simultaneous assessment of thousands of genes, often to identify differential expression across conditions. While sample size methods exist for static designs, few address time course experiments. This thesis introduces a flexible approach for sample size and power analysis in time course microarray studies using positive false discovery rate control. To account for non-normality and correlation in gene expression data, the method employs permutation-based null distributions for test statistics. Simulation results are compared with existing methods for both static and time course designs, demonstrating improved accuracy in error rate and sample size estimation.