The p value distributions of the six tests for the breast cancer data. The density

is not shown.

(a) (b)

a) The correlation between the negative base ten logarithm of the p values of

ests applied to the breast cancer data. (b) A Venn diagram of detected DEGs

S, the ORT, the COPA and the DOG models for the breast cancer data.

e 6.24 shows the expression distribution of the gene with the

n number 211966_at of the breast cancer data. The COPA p value

than 0.01, but the DOG p value was greater than 0.01. The gene

predicted as a DEG by the ORT model and the MOST model, but

test and the LSOSS model as well. The gene expression profile

mixed distribution of the normal and the cancer replicates. When

licates of both the normal and the cancer replicates were treated

tliers and were removed, the remaining two groups of replicates,

rmed a tight cluster, did not show a significant difference.