Gene-set distance analysis (GSDA): a powerful tool for gene-set association analysis

BMC Bioinformatics

Table 2 Results for of the pediatric AML analysis

Method	Chloroma	logWBC	EFS
GSDA (t-test)	0.020	\(< 10^{-29}\)	0.060
GSDA (perm)	0.029	\(<10^{-6}\)	0.051
GSEA	0.030	0.008	0.572
GSA	0.512	0.688	0.114
SAFE	0.561	0.014	0.336
GT	0.014	\(< 10^{-5}\)	0.250
TOTS	0.960	\(< 0.001\)	0.217
POST	0.027	\(< 10^{-5}\)	0.130
PIANO (Wilcoxon)*	0.804	0.004	0.209
PIANO (Fisher)*	0.0004	\(< 10^{-37}\)	0.092
PIANO (Stouffer)*	0.004	\(< 10^{-27}\)	0.164
PIANO (Reporter)*	0.742	0.004	0.288
PIANO (Tail Strength)*	0.744	0.016	0.239
PIANO (Mean)*	0.836	0.030	0.217
PIANO (Median)*	0.834	0.001	0.316
PIANO (Sum)*	0.858	0.042	0.221
PIANO (MaxMean)*	0.116	0.011	1.000
PIANO (GSEA)*	0.039	0.028	1.000
PIANO (FGSEA)*	0.014	0.029	1.000
PIANO (PAGE)*	0.044	0.096	0.112

The table shows the p values for each method for the association the expression of the KEGG AML pathway with clinical characteristics and survival in pediatric AML
*PIANO was performed with only the gene-level statistics with signifMethod=’nullDist’ for Wilcoxon, Fisher, Stouffer, and Reporter methods and signifMethod = ‘geneSampling’ and 999 permutations for the other methods

ISSN: 1471-2105