Skip to contents

Normalization of preprocessed data

Source: R/normalize.R
normalize.Rd

Apply a specified type of normalization to a data set.

Usage

normalize(dataSet, applyto = "sample", normalizeType = "quant", plot = TRUE)

Arguments

dataSet

The 2d data set of experimental values.

applyto

A string (default = "sample") specifying the target of normalization. Options are "sample" or "row" (across rows), and "protein" or "column" (across columns).

normalizeType

A string (default = "quant") specifying which type of normalization to apply:

  • "auto": Auto scaling (Jackson 1991) .

  • "level": Level scaling.

  • "mean": Mean centering.

  • "median": Median centering.

  • "pareto": Pareto scaling.

  • "quant": Quantile normalization (Bolstad et al. 2003) .

  • "range": Range scaling.

  • "vast": Variable stability (VAST) scaling. (Keun et al. 2003) .

  • "none": None.

plot

A boolean (default = TRUE) specifying whether to plot the boxplot for before and after normalization.

Value

A normalized 2d dataframe.

Details

Quantile normalization is generally recommended. Mean and median normalization are going to be included as popular previous methods. No normalization is not recommended. Boxplots are also generated for before and after the normalization to give a visual indicator of the changes.

References

Bolstad BM, Irizarry RA, Astrand M, Speed TP (2003). “A Comparison of Normalization Methods for High Density Oligonucleotide Array Data Based on Variance and Bias.” Bioinformatics, 19(2), 185–193. doi:10.1093/bioinformatics/19.2.185 .

Jackson JE (1991). A User's Guide to Principal Components. John Wiley \& Sons, New York, NY, USA. ISBN 9780471622673.

Keun HC, Ebbels TMD, Antti H, Bollard ME, Beckonert O, Holmes E, Lindon JC, Nicholson JK (2003). “Improved Analysis of Multivariate Data by Variable Stability Scaling: Application to NMR-based Metabolic Profiling.” Analytica Chimica Acta, 490(1–2), 265–276. doi:10.1016/S0003-2670(03)00094-1 .