PAA (version 1.7.1)

batchFilter.anova: Remove features which are differential regarding microarray batches / lots in a multi-batch scenario.

Description

Finds features which are differential regarding at least two microarray batches / lots in a multi-batch scenario (i.e., > 2 batches) via one-way analysis of variance (ANOVA) and removes them.

Usage

batchFilter.anova(elist=NULL, log=NULL, p.thresh=0.05, fold.thresh=1.5,
output.path=NULL)

Arguments

elist

EList or EListRaw object (mandatory).

log

logical indicating whether the data is in log scale (mandatory; note: if TRUE log2 scale is expected).

p.thresh

positive float number between 0 and 1 indicating the maximum Student's t-test p-value for features to be considered as differential (e.g., "0.5").

fold.thresh

float number indicating the minimum fold change for features to be considered as differential (e.g., "1.5").

output.path

string indicating a path for saving results (optional).

Value

An EList or EListRaw object without differential features regarding at least two microarray batches / lots.

Details

This function takes an EList or EListRaw object (see limma documentation) to find features which are differential regarding at least two microarray batches / lots in a multi-batch scenario (i.e., more than two batches). For this purpose, thresholds for p-values obtained from an one-way analysis of variance (ANOVA) and fold changes can be defined. To visualize the differential features a volcano plot is drawn. Then, differential features are removed and the remaining data are returned. When an output path is defined (via output.path) volcano plots and result files are saved on the hard disk.

Examples

cwd <- system.file(package="PAA")
load(paste(cwd, "/extdata/Alzheimer.RData", sep=""))
elist <- elist[elist$genes$Block < 10,]
elist <- batchFilter.anova(elist=elist, log=FALSE, p.thresh=0.001,
 fold.thresh=3)