MvsA.cpd {compdiagTools}R Documentation

Rank-vs-Residual Plots

Description

Uses hexbin to draw rank-vs-residuals plots with loess and running median average deviation.

Usage

MvsA.cpd(a.x, a.y, call = "", filename = NULL, sample.size = 10000, 
         win.size = 1000, win.step = trunc(length(a.x)/50), ...)

Arguments

a.x values of considered chip.
a.y values of reference chip, e.g. median over all chips.
call whether the data is from normalization of summary.
filename writes a bitmap/pdf if this argument is specified. The filename is completed with a file extension according to the type required and a postfix according to call.
sample.size How many points are sampled for running MAD
win.size how large is the average window for running MAD
win.step step size for running MAD
... further argument passed to 'bitmap'

Details

Uses hexbin to draw a density plot rather than a scatter plot. We expect equally distributed residuals for good quality chips. Therefore, the blue dots representing median average deviation from residuals of a running window should be located on a horizontal line. The read line represents a loess fit and whould be close to zero for the whole x-range.

Note

Use the additional argument type="pdfwrite" to generate PDF.

Author(s)

Stefan Bentink

See Also

bitmap

Examples

library(vsn)
library(golubEsets)
data(golubTest)

expr.mat <- exprs(vsn(golubTest, describe.preprocessing=FALSE))
expr.med <- apply(expr.mat, 1, median)
MvsA.cpd(expr.mat[,1], expr.med)

[Package compdiagTools version 1.5.3 Index]