splineDiffExprs: Differential expression analysis based on natural cubic...

In splineTCDiffExpr: Time-course differential gene expression data analysis using spline regression models followed by gene association network reconstruction

Description

The function compares time dependent behaviour of genes in two different groups. Applying empirical Bayes moderate F-statistic on differences in coefficients of fitted natural cubic spline regression models, differentially expressed in time genes are determined. The function is a wrapper of other R-functions to simplify differential expression analysis of time-course data.

Usage

splineDiffExprs(eSetObject, df, cutoff.adj.pVal=1, reference, intercept=TRUE)

Arguments

eSetObject	ExpressionSet object of class ExpressionSet containing log-ratios or log-values of expression for a series of microarrays
df	number of degrees of freedom
cutoff.adj.pVal	Benjamini-Hochberg adjusted p-value cut-off
reference	character defining which treatment group should be considered as reference
intercept	if TRUE, F-test includes all parameters; if FALSE, F-test includes shape parameters only; default is TRUE

Details

The function fits a temporal trend using a natural cubic spline regression to simulate nonlinear behaviour of genes over time.

The input eSetObject must be provided as an object of class ExpressionSet which contains SampleName, Time, Treatment and if applicable Replicates variables (columns) included in the phenotypic data of the eSetObject (pData(eSetObject)). Two types of Treatment defining two groups to compare have to be definied.

Replicates are not required. The time points for compared treatment groups should be identical.

User has to define number of degrees of freedom (df) for the spline regression model. Choosing effective degrees of freedom in range 3-5 is reasonable.

Time dependent differential expression of a gene is determined by the application of empirical Bayes moderate F-statistics on the differences of coefficient values of the fitted natural cubic spline regression models for the same gene in the two compared treatment groups. In other words, comparing the coefficient values of the fitted splines in both groups allows the detection of differences in the shape of the curves, which represent the gene expressions changes over time. Ouptut table containing Benjamini-Hochberg adjusted p-value (adj.P.Value) is used to define differentially expressed genes. The default value for cutoff.adj.pVal is set to 1, which means that all genes are included in output table.

Value

A data.frame with rows defining names/IDs of differentially expressed genes and additional columns described below.

The first columns contain all feature data of the eSetObject (fData(eSetObject)), if any feature data were defined. Otherwise, only one column row_IDs, containing the row names is created. The b_0, b_1,..., b_m coefficients correspond to the reference model parameters. The d_0, d_1,..., d_m coefficients represent the differences between the reference model parameters and the model parameters in the compared group. AveExprs refers to the average log2-expression for a probe (representing a gene) over all arrays. The F column contains moderate F-statistics, P.Value raw p-value and adj.P.Value Benjamini-Hochberg adjusted p-value.

Author(s)

Agata Michna

See Also

limma

Examples

## load "eSetObject" containing simulated time-course data
data(TCsimData)
pData(TCsimData)

## define function parameters 
df <- 3
cutoff.adj.pVal <- 0.01
reference <- "T1"
intercept <- TRUE

diffExprs <- splineDiffExprs(eSetObject = TCsimData, df, cutoff.adj.pVal, reference, intercept)
head(diffExprs,3)

splineTCDiffExpr documentation built on Oct. 5, 2016, 4:28 a.m.