Figure 2

Scatter plots of standard deviation versus mean log-intensity for BWH, NKI and EXPO-breast data sets, respectively. The shading codes the gene stability scores, with darker colors indicating higher scores. These three data sets are from different microarray platforms. The light gray points indicate the discarded genes (those with mean expression level below the β value – see Eq. 1). The curves correspond to equal score levels and are one score unit apart.