qualVarAnalysis: Tests association between qualitative variables and components.

Description

This function tests if the groups of samples formed by the variables are differently distributed on the components, in terms of contribution value (i.e of values in matrix A(icaSet)). The distribution of the samples on the components are represented using either density plots of boxplots. It is possible to restrict the tests and the plots to a subset of samples and/or components.

Usage

qualVarAnalysis(params, icaSet, keepVar, keepComp = indComp(icaSet), keepSamples = sampleNames(icaSet), adjustBy = c("none", "component", "variable"), method = "BH", doPlot = TRUE, typePlot = "density", addPoints = FALSE, onlySign = TRUE, cutoff = params["pvalCutoff"], colours = annot2col(params), path = "qualVarAnalysis/", filename = "qualVar", typeImage = "png")

Arguments

params

An object of class MineICAParams providing the parameters of the analysis.

icaSet

An object of class IcaSet.

keepVar

The variable labels to be considered, must be a subset of varLabels(icaSet).

keepComp

A subset of components, must be included in indComp(icaSet). By default, all components are used.

keepSamples

A subset of samples, must be included in sampleNames(icaSet). By default, all samples are used.

adjustBy

The way the p-values of the Wilcoxon and Kruskal-Wallis tests should be corrected for multiple testing: "none" if no p-value correction has to be done, "component" if the p-values have to be corrected by component, "variable" if the p-values have to be corrected by variable

method

The correction method, see p.adjust for details, default is "BH" for Benjamini & Hochberg.

doPlot

If TRUE (default), the plots are done, else only tests are performed.

addPoints

If TRUE, points are superimposed on the boxplot.

typePlot

The type of plot, either "density" or "boxplot".

onlySign

If TRUE (default), only the significant results are plotted.

cutoff

A threshold p-value for statistical significance.

colours

A vector of colours indexed by the variable levels, if missing the colours are automatically generated using annot2Color.

path

A directory _within resPath(params)_ where the files containing the plots and the p-value results will be located. Default is "qualVarAnalysis/".

typeImage

The type of image file to be used.

filename

The name of the HTML file containing the p-values of the tests, if NULL no file is created.

Value

Returns A data.frame of dimensions 'components x variables' containing the p-values of the non-parametric tests (Wilcoxon or Kruskal-Wallis tests) wich test if the samples groups defined by each variable are differently distributed on the components.

Details

This function writes an HTML file containing the results of the tests as a an array of dimensions 'variables * components' containing the p-values of the tests. When a p-value is considered as significant according to the threshold cutoff, it is written in bold and filled with a link pointing to the corresponding plot. One image is created by plot and located into the sub-directory "plots/" of path. Each image is named by index-of-component_var.png. Wilcoxon or Kruskal-Wallis tests are performed depending on the number of groups of interest in the considered variable (argument keepLev).

Examples

Run this code

## load an example of IcaSet
data(icaSetCarbayo)

## build MineICAParams object
params <- buildMineICAParams(resPath="carbayo/")

## Define the directory containing the results
dir <- paste(resPath(params), "comp2annot/", sep="")

## Run tests, make no adjustment of the p-values,
# for variable grade and components 1 and 2,
# and plot boxplots when 'doPlot=TRUE'.
qualVarAnalysis(params=params, icaSet=icaSetCarbayo, adjustBy="none", typePlot="boxplot",
                keepVar="GRADE", keepComp=1:2, path=dir, doPlot=FALSE)