Methods implementing Lattice displays for flow data
Description
Various methods implementing multipanel visualizations for flow data using
infrastructure provided in the lattice package. The original generics for
these methods are defined in lattice, and these S4 methods (mostly) dispatch
on a formula and the data
argument which must be of class
flowSet
or flowFrame
. The formula has to be fairly basic:
conditioning can be done using phenodata variables and channel names (the
colnames
slot) can be used as panel variables. See examples below for
sample usage.
Usage
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15  ## S4 method for signature 'formula,flowSet'
levelplot(x, data, xlab, ylab, as.table = TRUE,
contour = TRUE, labels = FALSE, n = 50, ...)
## S4 method for signature 'formula,flowSet'
qqmath(x, data, xlab, ylab, f.value = function(n)
ppoints(ceiling(sqrt(n))), distribution = qnorm, ...)
## S4 method for signature 'flowFrame,missing'
parallel(x, data, reorder.by = function(x)
var(x, na.rm = TRUE), time = "Time", exclude.time = TRUE, ...)
## S4 method for signature 'formula,flowSet'
parallel(x, data, time = "Time",
exclude.time = TRUE, filter = NULL, xlab = NULL, ylab = NULL, ...)

Arguments
x 
a formula describing the structure of the plot and the variables to be used in the display. 
data 
a 
xlab,ylab 
Labels for data axes, with suitable defaults taken from the formula 
as.table,contour,labels 
These arguments are passed unchanged to the corresponding methods in lattice, and are listed here only because they provide different defaults. See documentation for the original methods for details. 
n 
the number of bins on each axis to be used when evaluating the density 
f.value,distribution 
number of points used in QQ plot, and the
reference distribution used. See 
reorder.by 
a function, which is applied to each column. The columns
are ordered by the results. Reordering can be suppressed by setting this to

time 
A character string giving the name of the column recording time. 
exclude.time 
logical, specifying whether to exclude the time variable from a scatter plot matrix or parallel coordinates plot. It is rarely meaningful not to do so. 
filter 
flowCore filter 
... 
more arguments, usually passed on to the underlying lattice methods. 
Details
Not all standard lattice arguments will have the intended effect, but many should. For a fuller description of possible arguments and their effects, consult documentation on lattice (Trellis docs would also work for the fundamentals).
Methods
 qqmath
signature(x = "formula", data = "flowSet")
: creates theoretical quantile plots of a given channel, with one or more samples per panel levelplot
signature(x = "formula", data = "flowSet")
: similar to thexyplot
method, but plots estimated density (usingkde2d
) with a common zscale and an optional color key. parallel
signature(x = "flowFrame", data = "missing")
: draws a parallel coordinates plot of all channels (excluding time, by default) of aflowFrame
object. This is rarely useful without transparency, but that is currently only possible with thepdf
device (and perhaps the aqua device as well).
Examples
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28  data(GvHD)
qqmath( ~ `FSCH`  factor(Patient), GvHD,
grid = TRUE, type = "l",
f.value = ppoints(100))
## contourplot of bivariate density:
require(colorspace)
YlOrBr < c("#FFFFD4", "#FED98E", "#FE9929", "#D95F0E", "#993404")
colori < colorRampPalette(YlOrBr)
levelplot(asinh(`SSCH`) ~ asinh(`FSCH`)  Visit + Patient, GvHD, n = 20,
col.regions = colori(50), main = "Contour Plot")
## parallel coordinate plots
parallel(GvHD[["s6a01"]])
## Not run:
## try with PDF device
parallel(GvHD[["s7a01"]], alpha = 0.01)
## End(Not run)
