pdw: Plot density of each column in a dataset, using only one well...
In rxmenezes/rscreenorm: Normalization of Multiple Genetic Screens Data

Description Usage Arguments Value See Also Examples

Computes a kernel density per column in a dataset and makes a single plot with appropriate limits of all densities together. The graph also includes a legend by default.

pdw(data_rscreen, wtype = "sample", n = 512, mycols = NULL,
  na.rm = TRUE, mytitle = paste("Data density", wtype, "wells"),
  sel.reps = scr.names(data_rscreen), rescale = FALSE,
  scale.fun = c("asinh", "log2"), mytype = "solid",
  myxlab = "observed values", add.leg = TRUE, myleg = NULL,
  leg.side = c("right", "left"), lcex = 1, xlim = NULL)

`data_rscreen`	typically an rscreen.object, from which the columns of the 'data_only' slot are to be made density plots of. Otherwise a numeric matrix or a data frame with only numeric columns, in which case density plots are made of their columns. In the latter case, all entries (rows) are assumed to be of type 'sample'.
`wtype`	string indicating the well type to be used. Must be a label that exists in the wtype_ann slot.
`n`	the number of knots at which the density is to be computed - see `density` for details.
`mycols`	colours to be assigned to the lines of the densities computed. If left empty, `rainbow` is used to produce one colour per column in mydata.
`na.rm`	logical - should NAs per column be removed? Its value is passed on to the call to `density`.
`mytitle`	string containing the title to be used.
`sel.reps`	either a character vector containing labels for replicates to be used, or indices of columns from the 'data_only' slot of the data_rscreen object to be used. The default is to use all replicates in the 'data_rscreen' object.
`rescale`	logical, indicating whether or not the data is to be rescaled. Defaults to `FALSE`. If `TRUE`, the screen data will be rescaled per replicate using the function indicated in 'scale.fun'.
`scale.fun`	the name of the function to be used to rescale the data. Currently the only choices are 'asinh' for the hyperbolic-arc sine transformation, and 'log2' for the log2 transformation. Note that the hyperbolic-arc sine is equivalent to the logarithm for intermediate and high values, and to a linear transformation for low values. This transformation allows for zeros, which is a useful property when transforming count data. Ignored if 'rescale' is `FALSE`.
`mytype`	string indicating the line type to be used. If a single value is given, it is recycled across all lines (one for each column in mydata). If a vector is given, it is expected to be as long as the number of columns in mydata. The default is to use solid lines for all densities. For a list of all possible line types, see graphical parameter lty in the help file for `par`.
`myxlab`	string containing the legend to be used for the x-axis.
`add.leg`	logical. If `TRUE`, adds a legend to the plot.
`myleg`	the legend text to be shown. If left empty, and add.leg is `TRUE`, the column names of mydata are used.
`leg.side`	character, the side of the plot where the legend is to be placed. Defaults to 'right', with the other option being 'left'.
`lcex`	cex parameter for legend, if used. Must be numeric.
`xlim`	a numerical vector of length 2 giving the limits to be used for the x axis. If left empty, the range of values in mydata is used.

A plot including one density for each column of values in data_rscreen, using plot limits that allow for a complete displ)ay of each density, by default.

density, rainbow

mydata <- cbind(replicate(5, rnorm(500)), replicate(5, rnorm(500, mean=1)))
colnames(mydata) <- paste("Data", 1:10)
data.ann <- data.frame(geneID = paste("G", 1:500), wtype = rep("sample", 500))
data.screen <- list(data_ann = data.ann, data_only = mydata)
class(data.screen) <- "rscreen.object"
pdw(data.screen, lcex = .8)