PSEA (version 1.6.0)

swlm: Stepwise model selection

Description

Simple wrapper around stepAIC() (package MASS) to repeatedly perform stepwise model selection by AIC on several dependent variables (or responses, taken as rows of a matrix).

Usage

swlm(y,subset=NULL,upper,lower=formula(~1),direction='both',trace=FALSE,keep=NULL,verbose=FALSE)

Arguments

y

Numeric matrix (with responses as rows and samples as columns) or ExpressionSet. Typically the expression data with transcripts (i.e. for a microarray, probes) as rows and samples as columns. If an ExpressionSet is provided the expression data is extracted with the function exprs.

subset

Integer vector. Represents a subset of samples (specified as column indices in y) to use for model fitting. By default all samples are used.

verbose

logical. If TRUE (default) the response number being fitted is printed.

upper

see ?stepAIC

lower

see ?stepAIC

direction

see ?stepAIC

trace

see ?stepAIC

keep

see ?stepAIC

Value

swft

List of stepwise-selected models (see ?stepAIC)

Details

The initial model for the stepwise approach only contains an intercept term.

References

Kuhn A, Kumar A, Beilina A, Dillman A, Cookson MR, Singleton AB. Cell population-specific expression analysis of human cerebellum. BMC Genomics 2012, 13:610.

See Also

marker,lmfitst.

Examples

## Load example expression data (variable "expression")
## and phenotype data (variable "groups")
data("example")

## Four cell population-specific reference signals
neuron_probesets <- list(c("221805_at", "221801_x_at", "221916_at"),
		"201313_at", "210040_at", "205737_at", "210432_s_at")
neuron_reference <- marker(expression, neuron_probesets)

astro_probesets <- list("203540_at",c("210068_s_at","210906_x_at"),
		"201667_at")
astro_reference <- marker(expression, astro_probesets)

oligo_probesets <- list(c("211836_s_at","214650_x_at"),"216617_s_at",
		"207659_s_at",c("207323_s_at","209072_at"))
oligo_reference <- marker(expression, oligo_probesets)

micro_probesets <- list("204192_at", "203416_at")
micro_reference <- marker(expression, micro_probesets)

## Stepwise model selection for 2 transcripts (202429_s_at and 200850_s_at)
## and focusing on control samples (i.e. groups == 0)
swlm(expression[c("202429_s_at", "200850_s_at"),],
	subset = which(groups == 0), 
	upper = formula(~neuron_reference + astro_reference +
			oligo_reference + micro_reference))