Figure 3

Regression model. The percent variance explained by each of the variables and their cross terms (second-order interactions). The black bars show significant terms (p < 0.001). The variables are the following: data set - the six data sets Alizadeh, Finak, Galland, Herschkowitz, Jones and Ye; norm - the five normalizations no.norm, norm.pt, norm.pt.bkg, norm.glob and norm.glob.bkg; standard - standardization or not; MV - missing value imputation using one of the methods ROW or SVD; GSnoofgenes - gene selection method and number of selected genes (in total 13 variants); method - the clustering methods hierarchical clustering (six different settings), k-means, PAM (two variants), SOM and Mclust. A cross term between A and B is denoted A:B. The percent variability explained by the residuals is the variability not explained by the regression model.