Fig. 8
From: Integrating mean and variance heterogeneities to identify differentially expressed genes
Deflations due to the over adjustment of the experiment-wide data structure. Among all the 79 global PCs, only PC4 displayed significant mean heterogeneity (p WT =4.49E-15). PC6, 9, 14, 28, 38, 49 and 78 displayed variance heterogeneity (p LF ranged from 3.18E-4 to 0.0419). After adjusting for the significant global PCs, age and menopausal status, the Q-Q plots of all the five competitors displayed severe deflations. All the genomic inflation factors turned out to be much smaller than 1. The Q-Q plots of the four mean heterogeneity tests fell below the diagonal, where those of the WT and the MWT fell below the lower limit of the 95% concentration band. Global PCs did not distinguish informative heterogeneities and impediment heterogeneities. The significant global PCs would account for big portions of informative mean and variance heterogeneities due to DE genes. Therefore, adjusting for the significant PCs of all the experiment-wide gene probes would reduce statistical powers