(a) (b)

he modified t test applied to the prostate cancer data. ‘rho’ stands for the

coefficient. (a) The correlation between raw (original) base two logarithm of

e and the estimated base two logarithm of fold change from a limma model. (b)

tion between the original t test p values and the SAM t test p values.

6.2 shows the confusion matrix for comparing the t test and the

t test results, where the critical p value was 0.01. This analysis

at two approaches still had some degrees of discrepancy and

6 shows the Venn diagram of this discrepancy analysis between

s discovered by the modified t test and the DEGs discovered by

Two tests generated results which are significantly different.

The confusion matrix to compare the t test and the modified t test results for the

ncer data.

Original t test

Non-DEG

DEG

SAM t test

Non-DEG

50393

1136

DEG

1377

1769

he Venn diagram of DEGs predicted by the t test and the modified t test for the

ncer data.