The mePipe R package provides a convenient way to carry out several steps that are often required as part
of an eQTL analysis. Matrix-eQTL is used
to carry out the association testing. This very fast and supports the use of many covariates without a
substantial increase in run time. In addition mePipe provides facilities to compute principle components
of the gene expression data and include a subset of these as covariates to account for confounding variation
and maximise the number significant associations. Some functionality to account for LD structure is also available.
All computations can either be carried out on a single processor or run in parallel (as far as possible) on an
SGE compute cluster.
The mePipe package requires R version 2.15 or newer to run and the following R packages have to be available. If they are not already present they need to be installed before mePipe:
The first four of these are available from CRAN and Bioconductor and can be installed from within R with
source("http://bioconductor.org/biocLite.R")
biocLite(c("MatrixEQTL", "trio", "optparse", "XML", "snow"))Unfortunately Rsge has recently been removed from CRAN.
The latest version is available from here. To install the latest version
directly from GitHub, use the following command from within R (requires the devtools package)
devtools::install_github("Rsge", "humburg")The latest version of the R source package is available here. Download this into your home directory and (on the machine on which you would like to add the R package) run
R CMD INSTALL mePipe_latest.tar.gzThe R package includes the runMatrixEQTL.R script that allows use of the
MatrixEQTL package from the command line.
By default this will be installed into <R_library>/mePipe/exec/. When installing mePipe for the first
time it is recommended to create a link to this script in a more convenient place, e.g.
mkdir ~/bin
cp -s <R_library>/mePipe/exec/runMatrixEQTL.R ~/bin/