Robust extraction of functional signals from gene set analysis using a generalized threshold free scoring function

Table 3 Rank and Pearson correlations for each method's results between the split parts of the dataset

	1st split		2nd split		3rd split		4th split
method	Rank	Pearson	rank	Pearson	rank	Pearson	rank	Pearson
GSZ-score	0.6878	0.8427	0.7021	0.8463	0.7051	0.8521	0.6867	0.8194
t-test	0.6236	0.7291	0.686	0.7715	0.6932	0.7763	0.7404	0.7996
KS test	0.5402	0.585	0.5768	0.6197	0.5824	0.6282	0.5845	0.6145
modKS	0.6183	0.6397	0.6443	0.6658	0.6343	0.6554	0.6952	0.7083
iGA	0.6134	0.7261	0.6424	0.7401	0.6358	0.7562	0.6239	0.7243

Rank and Pearson correlation between the results from two halves of the dataset for each method (case i). The best result is highlighted with bold font and the second best is underlined. Notice that GSZ-score has the highest correlation with both correlation measures. t-test shows the second best performance. The only deviation is the 4th split where t-test shows the best rank correlation, although even in that case GSZ-score still shows the best Pearson correlation.

ISSN: 1471-2105