This normalization option allows you to normalize sections of a sample rather than normalizing over the entire sample. This is especially important if you used multiple arrays for each experimental point or if there is some reason you need to normalize sections of an array separately from one another. Region normalization is not a separate mathematical formula the way the previous normalizations discussed in this chapter are. Using this normalization means if you normalize to negative controls, to positive controls or normalize each sample to itself you do not actually normalize over each sample, but rather perform the normalization over each region. Hence the formulas for these three normalization options become:
Normalizing to Negative Controls for a Region:
(the control strength of gene A in region Y of sample X)
-(the median signal of the negative controls in region Y of sample X)
Normalizing to Positive Controls for a Region:
(the control strength of gene A in region Y of sample X)
(the median signal of the positive controls in region Y of sample X)
Normalizing Each Region to Itself:
(the control strength of gene A in region Y of sample X)
(the median of all of the measurements taken in region Y of sample X)
See Experiment Normalizations for how to implement this normalization option from within GeneSpring and for how to define a region.