spillover: Compute a spillover matrix from a flowSet

Description

Spillover information for a particular experiment is often obtained by running several tubes of beads or cells stained with a single color that can then be used to determine a spillover matrix for use with compensate.

When matching stain channels in x with the compensation controls, we provide a few options. If ordered, we assume the ordering of the channels in the flowSet object is the same as the ordering of the compensation-control samples. IF regexpr, we use a regular expression to match the channel names with the filenames of the compensation controls. By default, we must "guess" based on the largest statistic for the compensation control (i.e., the row).

Usage

"spillover"(x, unstained = NULL, patt = NULL, fsc = "FSC-A",  ssc = "SSC-A", method = "median", stain_match = c("intensity", "ordered", "regexpr"), useNormFilt=FALSE, pregate = FALSE, plot = FALSE, ...)

Arguments

A flowSet of compensation beads or cells

unstained

The name of index of the unstained negative control

patt

An optional regular expression defining which parameters should be considered.

fsc

The name or index of the forward scatter parameter

ssc

The name or index of the side scatter parameter

method

The statistic to use for calculation. Traditionally, this has been the median so it is the default. The mean is sometimes more stable.

stain_match

Determines how the stain channels are matched with the compensation controls. See details.

useNormFilt

Logical Indicating whether to apply a norm2Filter first before computing the spillover

pregate

logical. Should we pregate each channel before computing the spillover matrix? By default, no.

plot

logical. Plots the kernel density for each channel when pregating. Displays the gate used. If pregate is set to FALSE, this argument is ignored.

...

Additional arguments passed to rangeGate.

Value

Details

The algorithm used is fairly simple. First, using the scatter parameters, we restrict ourselves to the most closely clustered population to reduce the amount of debris. The selected statistic is then calculated on all appropriate parameters and the unstained values swept out of the matrix. Every sample is then normalized to [0,1] with respect to the maximum value of the sample, giving the spillover in terms of a proportion of the primary channel intensity.

References