Quantile normalization

In order to get the same overall distribution of intensities, the array-intensity values of n arrays are normalized by projecting each quantile of intensities to lie along the unit diagonal. In n dimensions all n data vectors should have the same distribution such that plotting the normalized quan-tiles in n dimensions leads to the unit vector (1/Vn,...,1/Vn). To end up with the same distribution for all arrays, one takes the mean quantile and substitutes it with the value of the data point in the original dataset. Quantile normalization was proposed by Bolstad et al. (21) and is implemented in the affy-package in Bioconductor (http://www.bioconductor.org/; library: affy, function: normalize.quantiles and normalize.quantiles.robust).

0 0

Post a comment