normalizeAcrossSlides.MAData: Normalizes across slides
In HenrikBengtsson/aroma: An R Object-oriented Microarray Analysis package

Description Usage Arguments Author(s) See Also Examples

Normalizes across some or all slides. After doing within-slide normalization (see *normalizeWithinSlide(), an across-slide normalization must be performed before the data on the different slides can be compared with each. Across-slide normalization scales the log ratios (M) for all slide so each slide gets the same log-ratio deviation based on the robust deviation measure Maximum Absolute Deviation (MAD).

If one would like to normalize the deviation across different data set, i.e. different MAData objects, one can make use of the argument newMAD, which forces the slides to get a specific median absolute deviation value.

Note, in the case where one set of slides comes from one type of experimental setup and a second set of slides comes from another setup, and they are stored in the same MAData object, these two groups of slides can be normalized together using one (in other words, you do not have to normalize the two groups seperately and the rescale them with newMAD).

Also note that it is only the log ratios, M, are affected by the normalization, i.e. the log intensities, A, are not changed.

1 2	## S3 method for class 'MAData' normalizeAcrossSlides(this, slides=NULL, newMAD=NULL, ...)

`slides`	The set slides to be used and to be normalized. If `NULL` all slides will be normalized.
`newMAD`	After the normalization all slides will have an maximum absolute deviation of `newMAD`. If `NULL`, the result will be the same as if `newMAD` would be set to the geometrical mean of each individual slide's MAD.

Henrik Bengtsson (http://www.braju.com/R/)

For an detailed explanation of the robust deviation measure MAD (Median Absolute Deviation) see mad. For within-slide normalization see *normalizeWithinSlide(). For more information see MAData.

  # The option 'dataset' is used to annotate plots.
  options(dataset="sma:MouseArray")

  SMA$loadData("mouse.data")
  layout <- Layout$read("MouseArray.Layout.dat", path=system.file("data-ex", package="aroma"))
  raw <- RawData(mouse.data, layout=layout)
  ma <- getSignal(raw)

  layout(matrix(1:9, ncol=3, byrow=TRUE))

  # Plot data before within-slide normalization
  for (k in 1:3)
    boxplot(ma, groupBy="printtip", slide=k, main=paste("No normalization - #", k, sep=""))

  # Plot data after scaled print-tip normalization
  normalizeWithinSlide(ma, "s")
  for (k in 1:3)
    boxplot(ma, groupBy="printtip", slide=k, main=paste("Within-slide norm. - #", k, sep=""))

  # Plot data after across-slide normalization
  normalizeAcrossSlides(ma)
  for (k in 1:3)
    boxplot(ma, groupBy="printtip", slide=k, main=paste("Across-slide norm. - #", k, sep=""))