Fig. 3
From: Reproducible detection of disease-associated markers from gene expression data
Scores obtained using the t-statistic, Wilcoxon sum-rank statistic, and sign-sum statistic with two sampling sizes. From left to right, n=200 (upper), n=1000 (lower). Horizontal axis denotes the gene indices. Vertical label denotes the ranking score. High score indicates relative importance for discrimination of class labels. The first 100 genes are homo informative genes, the next 100 are the hetero informative genes, and the last 800 genes are the non-informative genes