Study Design in DIGE-Based Biomarker Discovery - Difference Gel Electrophoresis (DIGE): Methods and Protocols - page 205

Fig. 2. Two examples for the relationship between the difference in mean standardized
abundance ( D ) and the effect size ( q ). The effect size of control vs. group 1 is 4.0 and of
control vs. group 2 is 1.5. The graphs show the different degree of overlap (indicated in
gray ) of the distributions of expression values of the control and the respective group.

(independent or dependent) according to your study design.
Activate the Average Ratio and the Student's t test option and
defi ne the two experimental groups which should be com-
pared. Finally, select the FDR option.

7. The DeCyder software package expresses the degree of dispar-

ity in spot mean abundance between two experimental groups
as “average ratio” or “av R .” Other software packages and
many publications use the term “fold-change” for the same
value. av R derives from the ratio of the mean standardized
abundances (here termed R ¢):
(a)

R ¢= meanSA / meanSA .

1 c

(b) If R ¢ > 1, then av R = R ¢.
(c) If R ¢ < 1, then av 1 R R mean SA 1 or c mean stan-

dardized abundance of group-1 and control, respectively.
Note that the “average ratio” (or “fold-change”) should
not be confused with the effect size ( q ). Although both measures
are used to describe the disparity in expression of a protein,
only the effect size considers the standard deviation (and hence
the biological variation). Accordingly, the “average ratio” can-
not directly be converted to effect size.

=- ¢

8. According to the guidelines of the European (EMA) and the

American (FDA) regulatory agencies, the maximal allowed
technical variation of routine diagnostic methods (e.g., ELISA)
is 20% ( 14, 15 ). It seems therefore reasonable to include only
spots with an average ratio above ±1.2 in the validation
experiment.

9. If several statistical tests are being performed simultaneously,

controlling the FWER to adjust for multiple testing is more