Robust Significance Analysis of Microarrays by Minimum β-Divergence Method

Figure 6

Comparison of the top selected genes by different methods for the Colon, Leukemia, and Platinum Spike datasets. Colon dataset, Venn diagram of top 100 genes estimated by (a) the ANOVA, SAM, LIMMA, and proposed method or by (b) the KW, SAM, BRIDGE, and proposed method. Leukemia dataset, Venn diagram of top 100 genes estimated by (c) the ANOVA, SAM, LIMMA, and proposed method or by (d) the KW, SAM, BRIDGE, and proposed method. Platinum Spike dataset, (e) M-A plot for randomly selected small-sample size from 9 controls and 9 tests samples in presence of one outlier in 20% of 1944 valid DE genes. The red asterisk, blue triangle, and black circle are used for 312, 9, and 40 genes detected by proposed method, LIMMA, and SAM that are common with valid set (see Figure S5(c)).

(a)

(b)

(c)

(d)

(e)