These methods are applicable if only a single cDNA chip was used.
The data should be lowess-normalized before running this function. If the data has already been normalized using some other means, it need not to be done again.
Noise-envelope finds the genes that are within significance cut-off (number of standard deviations) from the chip mean. Newton returns the genes that have the bayesian posterior probability higher than the significance cut-off (log odds of being differentially expressed).
A list of genes with expression value.