To normalize in the context of DNA microarrays means to standardize your data to be able to differentiate between real (biological) variations in gene expression levels and variations due to the measurement process. Normalizing also scales your data so that you can compare relative gene expression levels.
GeneSpring assumes that the data that you have entered is raw data that needs to be normalized. Note that if your data has been pre-normalized around a median other than 1, it may not be interpreted accurately during analysis. If your data is pre-normalized this way, please refer to Use Constant Values or Normalizing Each Sample to a Hard Number.
There are several ways to normalize your data in GeneSpring. Typically, you will want to do either one per-chip normalization together with one per-gene normalization or one per-spot normalization with one per-chip normalization. There are important exceptions to this, which are discussed below under the relevant normalization.
Note also that the order in which normalizations are performed is mathematically significant; GeneSpring performs them in the order in which they are listed here (and in the Experiment Normalizations window).
To get to the Experiment Normalizations window to assign normalizations, select
Experiments > Experiment Normalizations
.
To estimate background noise, some chips come with negative control spots that do not correspond to mRNA from the species under study. Even if your imaging software automatically subtracts background fluorescence, you may still want to tell GeneSpring to normalize to negative controls. The formula used here is:
If you are conducting a two-color experiment, you will probably want to do a per-spot normalization. The formula for this normalization is:
(signal strength of gene A in sample X)
(control channel value for gene A in sample X)