Biology Reference
In-Depth Information
1040
1030
1020
A
C
D
E
F
1010
1000
990
980
1
2
3
4
A
Tissue samples
1.5
1
0.5
A
C
D
E
F
0
−
0.5
−
1
−
1.5
1
2
3
4
B
Tissue samples
FIGURE 12-9.
Effects of normalization. The data in panel A have been normalized, and the results plotted in
panel B.
samples, although gene C is expressed at a relatively higher level.
Genes E and F are also co-regulated, whereas gene D does not seem to
change expression much across samples. It would be desirable to group
genes based on similar expression patterns, so A and C should be
grouped together, although in the current situation C might be grouped
with D because of their similar average values across samples. In such
cases, normalizing the gene expressions in the expression matrix across
the samples would be appropriate.
This can be done by calculating the mean value and standard deviation
for each row of the matrix X, and then normalizing each entry of the
row by subtracting the mean value and dividing by the standard
deviation. For the entries in row 3, for example, the calculations are as
follows:
