normalizeD2 {DAPAR} | R Documentation |
Provides several methods to normalize data from a matrix. They are organized in four main families : Strong Rescaling, Median Centering, Mean Centering, Mean CenteringScaling. For the first family, two sub-categories are available : the sum by columns and the quantiles method. For the three other families, two categories are available : "Overall" which means that the value for each protein (ie line in the expression data tab) is computed over all the samples ; "within conditions" which means that the value for each protein (ie line in the matrix) is computed condition by condition.
normalizeD2(qData, labels, method, type, scaling = FALSE, quantile = 0.15)
qData |
A dataframe that contains quantitative data. |
labels |
A vector of strings containing the column "Label" of
the |
method |
One of the following : Global Alignment, Quantile Centering, Mean Centering. |
type |
For the method "Global Alignment", the parameters are: "sum by columns": operates on the original scale (not the log2 one) and propose to normalize each abundance by the total abundance of the sample (so as to focus on the analyte proportions among each sample). "Alignment on all quantiles": proposes to align the quantiles of all the replicates; practically it amounts to replace abundances by order statistics. For the two other methods, the parameters are "overall" (shift all the sample distributions at once) or "within conditions" (shift the sample distributions within each condition at a time). |
scaling |
A boolean that indicates if the variance of the data have to be forced to unit (variance reduction) or not. |
quantile |
A float that corresponds to the quantile used to align the data. |
A matrix normalized
Samuel Wieczorek, Thomas Burger
require(DAPARdata) data(Exp1_R25_pept) qData <- Biobase::exprs(Exp1_R25_pept[1:1000]) labels <- Biobase::pData(Exp1_R25_pept[1:1000])[,"Label"] normalizeD2(qData, labels, "Quantile Centering", "within conditions", quantile = 0.15)