
Data preprocessing and n ormalization 741
different, while the plot shown here in Fig. 20.15, as well as the coefficient of
variance plot in Fig. 19.17 in Chapter 19 did show this array as being different.
> boxplot(rawsum.lmT, col=mycol, main="VST transformed")
> boxplot(rawsum.lmQTL, col=mycol, main="quantile")
> boxplot(rawsum.lmVSN, col=mycol, main="vsn")
> boxplot(rawsum.lmRIV, col=mycol, main="rankinvariant")
The box plots of the same data are shown in Fig. 20.20. In this figure, the
top left panel s hows the VST transformed data, which is used as the input for
all three normalization methods. As noted in the intensity distribution plot,
the array