peakNorm: Normalize batch differences in intensities by aligning peaks
In flowSpecs: Tools for processing of high-dimensional cytometry data

Description Usage Arguments Value Examples

This function is intended to be used on standardized controls, preferrably one standard that has been acquired with every batch. This function needs to be applied separately for each batch, with the same standard.

peakNorm(
  fs,
  ctrlPos,
  standFF,
  transNames = FALSE,
  transCoFacs,
  volThresh = 0.005,
  normOrNot = rep(TRUE, ncol(standFF))
)

`fs`	The flowSet to be normalized
`ctrlPos`	The position in the flowSet that contains the internal control
`standFF`	The external standard to which all batches should be normalized
`transNames`	If parameters are not transformed prior to running this, internal transformation is necessary to identify peaks, so specify which variables that need transformation here. The data for these variables are untransformed at the end, so the data will have the same scale in and out.
`transCoFacs`	Also inherited from arcTrans. Low values (2-10) for CyTOF data, high values (200-2000) for flow data, all depending on the number of input channels.
`volThresh`	The threshold for how small the area of a peak can be compared to the largest peak, and still count.
`normOrNot`	This needs to be a logical vector with the same length as the number of columns in the fs and standFF. If it is known that a certain variable should not be normalized, that information can be specified here.

A normalized flowSet. In addition, output to console, to clarify #for which markers the same number of peaks was detected in the control and the standard, as normalization will only be applied for these.

#Load uncompensated data and spectral matrix.
data(fullPanel)
data(specMat)

# And now unmix
fullPanelUnmix <- specUnmix(fullPanel, specMat)

 #Create a new file with the value 1000 added to all values
 library(flowCore)
 fullPanelPlus1000 <- flowFrame(exprs(fullPanelUnmix)+1000)
# Check how they differ.
range(exprs(fullPanelUnmix)[,1])
# 143 1187733
range(exprs(fullPanelPlus1000)[,1])
# 1143 1188733

#Now normalize the new one to the old. NB! Here we will only
normPanel1000 <- peakNorm(flowSet(fullPanelPlus1000), 1, fullPanelUnmix,
transNames = colnames(fullPanelUnmix)[6:18], transCoFacs = 500)

#And now check the new result
range(exprs(normPanel1000[[1]] )[,1])
# 143 1187733