correction [26]. Compared with other processes, the results of
different normalization methods are significantly different, and
should be compared before making decision.
Figure6 shows the box plots of the arrays after background
correction byrmamethod and normalization by three different
methods. Compared with the other two normalization methods,
thequantilenormalization method gives better results.Fig. 5Intensities and qualities of cleaner data. (a) R code for showing the intensity and quality control of
cleaner data. (b) The intensities of 12 arrays are shown in box plots. (c) A summary of quality control metrics
for 12 arrays
Building Biological Networks from Gene Expression Data 185