limma.trans.running: Transformation method with limma.

In lunge111/intSEQ: Differential Expression Analysis of RNA-seq Data with Integrated Likelihood Method

Description

Implement limma trans. See detail for default options.

Usage

limma.trans.running(count.data, cond, normalize= TRUE)

Arguments

count.data	A numeric matrix of the count table.
cond	A vector of factors or numeric of the predictor variable.
normalize	Logical, whether to use "TMM" normalization.

Details

The limma trans first transform the data with the function log(y + sqrt(y^2 + y/θ^2) + 1/(2 * θ^2)), then estimate the θ such that the regression line of variance on started log of the mean has zero slope.

For the transformed data, we use calcNormFactors in edgeR to calculate the normalization factors with method equals to "TMM". Then we rescale each column by the 10^6 divided by the product of normalization factor and library size. It is a "TMM" version of count per million. Then we call lmFit and eBayes functions in limma with default options.

Value

An object of class MArrayLM

Author(s)

Yilun Zhang, David Rocke

References

Our first paper.

See Also

tran1, tran.est, tran.reg

Examples

data(count.data)
data(condition)
pv.trans <- limma.trans.running(count.data,condition)

lunge111/intSEQ documentation built on May 20, 2019, 9:38 a.m.