Skip to main content
Fig. 15 | BMC Systems Biology

Fig. 15

From: DGCA: A comprehensive R package for Differential Gene Correlation Analysis

Fig. 15

Performance comparison of DGCA/MEGENA, DICER and DiffCoEx in detecting differential correlation modules. In the simulation study, there were two modules and 30 genes were assigned to each module, within which a fraction (κ) of the gene pairs had differential correlation, where κ varied from 0.5 to 1 across simulation runs. We also simulated different sample sizes in each condition (n = 100, 200, 300, and 400). The highest sensitivity (a; dots = means, lines = standard errors of the mean) and Jaccard index b between the detected and true modules were calculated and averaged for each method (black for DGCA/MEGENA, blue for DICER and red for DiffCoEx) across all the independent runs in the simulation study. In the case that no module was detected by a method, all the genes were considered to belong to one single module for calculating Jaccard index, indicated by the dotted horizontal line. Each Jaccard index mean was calculated from ten simulation runs and two true modules per run

Back to article page