FIG. 6.

The number of unrelated E-genes might effect the power of the testing: Rows correspond to the number of unrelated E-genes in the simulated data sets (0, 100, 500, 1000), while columns represent true relations. In each plot, the x-axis shows the different noise levels while the y-axis shows the relative frequency of rejecting the different relations, which are marked by different colors (R1: blue, R2: purple, R3: green, R4: red, and R5: orange). Results show that the number of unrelated E-genes effect the power of the test. We still have hardly any false-positive rejections but gain power in rejecting the incorrect relations by decreasing the number of E-genes with (0,0) pattern.