Figure 1From: RCMAT: a regularized covariance matrix approach to testing gene setsNull distribution p-values. Null distribution p-values for both the RCMAT and the method of Kong et al. under the null hypothesis (no average difference). P-values for 100 simulated data sets within each of six conditions are given. Along the vertical axis is the expected p-value under the assumption of no difference between the two phenotypes; on the horizontal axis is the corresponding actual p-value obtained from the simulation.Back to article page