PREDA (version 1.18.0)

SODEGIR_GEstatistics: Wrapper function for gene expression statistics preprocessing for SODEGIR analysis

Description

Wrapper function for gene expression statistics preprocessing for SODEGIR analysis.

Usage

# SODEGIR_GEstatistics(.Object, pData_classColumn=NULL,
# referenceGroupLabel=NULL,
# statisticType=c("tstatistic", "FC", "FCmedian", "eBayes", "SAM"),
# singleSampleOutput=TRUE, varianceAll=FALSE)
SODEGIR_GEstatistics(.Object, ...)

Arguments

.Object

An object of class ExpressionSet containing gene expression input data

...

See below

pData_classColumn:

Column of phenoData slot from the ExpressionSet object, containing the label of sample classes

referenceGroupLabel:

Specify which class label is used for the reference sample used in computing statistics for differential expression.

statisticType:

Stastistic for differential expression that is computed on input data. Possible values are "tstatistic", "SAM" (SAM statistical score for differential expression), "FC" (Fold Change), "FCmedian" (fold change computed on medians)

singleSampleOutput:

Logical, if TRUE a statistic comparing each sample with the reference group is computed.

varianceAll:

This parameter affect the computation only when singleSampleOutput is TRUE. varianceAll is itself a logical parameter. If TRUE, all pathological (e.g. tumor) samples and all normal (reference) samples are used to estimate variance in the comparison of individual pathological samples to the normal reference, as described in the original SODEGIR apper by Bicciato et al. (Nucleic Acids Res. 2009).

The original SODEGIR statistic for Gene Expression was based on the SAM score. Therefore in the current PREDA version the varianceAll=TRUE parameter can be used only for SAM statistic: when singleSampleOutput is TRUE and a different statisticType is used, the variance is actually computed using only the normal (reference) samples.

If FALSE (default value), the computation of statistics for single sample VS reference comparisons only take into account the variance in the reference group of samples.

Value

The output is returned as a matrix.

Details

Using an ExpressionSet object as input, statistics for differential expression are computed comparing each sample with the reference group.

References

Silvio Bicciato, Roberta Spinelli, Mattia Zampieri, Eleonora Mangano, Francesco Ferrari, Luca Beltrame, Ingrid Cifola, Clelia Peano, Aldo Solari, and Cristina Battaglia. A computational procedure to identify significant overlap of differentially expressed and genomic imbalanced regions in cancer datasets. Nucleic Acids Res, 37(15):5057-70, August 2009.

See Also

preprocessingGE, SODEGIRpreprocessingGE, ExpressionSet