pipeLIMMA: A pipeline for LIMMA.

In jtlovell/RNAseqDE: A pipeline for differential expression analyses in R

Description

pipeLIMMA Run a pipeline of LIMMA functions for differential gene expression.

Usage

pipeLIMMA(counts, info, formula = NULL, contrast.matrix = NULL,
  block = NULL, design = NULL, runVoom = TRUE,
  use.qualityWeights = TRUE, use.topTable = FALSE, geneIDs = NA,
  verbose = TRUE, plotVoom = FALSE, ...)

Arguments

counts	A count matrix
info	An experimental design matrix
formula	A character string that can be coerced to a formula. Specify if a contrast model is not desired.
contrast.matrix	A matrix of contrasts, usually created by a call from limma:makeContrasts. Used only when limma:contrast.fit is intended as the statistical modeling function. NULL values force a traditional test of effects via limma::lmFit/ebayes.
block	A string that represents an individual that was repeatedly measured, if NULL, runs the analysis without a blocking / duplicate correlation factor
design	A design matrix, usually created by a call from "model.matrix". Used only if limma:contrast.fit is intended as the statistical modeling function. NULL values force a traditional test of effects via limma::lmFit/ebayes.
runVoom	Logical, if TRUE, normalizes the counts matrix via voom. If FALSE, assumes the counts matrix is already voom-normalized. Pre-running voom will speed up analyses with multiple pipeLIMMA calls.
use.qualityWeights	Logical, run voom with quality weights or not?
use.topTable	Logical, report F-statistics across all factors? If true, a third element is returned called fstats.
geneIDs	The names of genes. If NA, use row names from counts matrix
verbose	Logical, return progress updates?
plotVoom	Logical, plot the voom fit? Defaults to FALSE
...	additional arguments, not currently in use.

Details

This function runs the following pipeline:

• 1. calculate normalization factors via edgeR::calcNormFactors

• 2. Run limma::voom transformation

• 3. Run limma:lmFit linear modeling

• 4. Run limma::ebayes statistical modeling

• 5. Collect Log2 Fold-Changes using limma:topTable

• 6. Collate and ouput statistics and voom transformed data

Value

a list with 2 elements (if simple=TRUE)

• "stats": the statsistics generated from ebayes and topTable

• "voom": the voom normalized counts data

• "fstats": if a formula is provided, the toptable returned fstatsistics across all estimates of each factor

Examples

## Not run: 
data(kidney) # from simseq
counts<-kidney$counts
counts<-counts[sample(1:nrow(counts),1000),]
info<-data.frame(rep=kidney$replic,
                 treatment=kidney$treatment)
stats<-pipeLIMMA(counts=counts,
                 info=info,
                 formula = " ~ treatment",
                 block=info$rep)
stats<-pipeLIMMA(counts=counts,
                 info=info,
                 formula = " ~ treatment",
                 block=NULL)

## End(Not run)

jtlovell/RNAseqDE documentation built on May 10, 2019, 8:06 a.m.