Detection of differentially regulated genes: Differential Co-Expression, Stouffer Merge and Time Course Analysis

Motivation: To detect sets of differentially co-expressed genes in two phenotypically distinct sets of expression profiles, to transform a set of p-values scoring the significance of probeset fold changes to a set of gene p-values using Stouffer’s method and to provide a platform for the analysis of time course data. Results: A greedy approach to find differentially co-expressed genes was implemented and validated on simulated data. Applied to real data the method was found far to unreliable. Stouffer’s method was implemented and applied to two clinical datasets with reliable outcome. Two distinct time course analysis methods as provided by the Bioconductor project were integrated and compared to each other showing an enormous discrepancy refering to the resulting toplist of genes of each method. Availability: All methods were implemented in R and integrated to the WFS (workflow system). Hence the R programs and the belonging WFS transitions may be obtained from the author.