Time course gene microarray is an important tool to identify genes with differential expressions over time under different conditions. The traditional analysis of variance type of longitudinal investigation may not be applicable, however, because the experimental designs for time course gene expression data cannot always be consistent across subjects, e.g., sampling  rates and time intervals among different subjects may not be the same. Moreover, missing data is very common due to contamination in microarray experiments. Here we use functional principal components in covariance analysis to test hypotheses in the change of mean curves. A permutation test under a mild assumption is used to make the method more robust. We illustrate the method by a simulation study and a real data analysis. The simulation study suggests that the proposed method outperforms the recently proposed EDGE and a two-way mixed effects ANOVA model under reasonable gene expression models. The proposed method is further illustrated by an application to transcriptional profiles of blood cells microarray experiment

of treated and untreated individuals.