Downloading and installing the package
Multiseq depends on ashr, a package in development that you can download and install following instructions here. It also depends on Rcpp, and therefore it requires an R version higher or equal to 3.0.0. To install the package, install ashr and then type the following commands in R:
#in orded to install rhdf5 you have to use the Bioconductor repository (tune parameter repos in install.packages) biocLite("rhdf5") install.packages("path/to/multiseq.tar.gz",repos=NULL,type="source")
where you should replace "path/to/multiseq.tar.gz" with the path to the downloaded multiseq package.
Some optional steps (only working in UNIX/LINUX) are needed if you want to use multiseq on sequencing data in hdf5, bam, bigWig, or bigBed format or if you want to visualize input/output sequencing data in the UCSC Genome Browser, i.e., if you want to run functions
Optional steps to:
Read data in hdf5, bam, or bigWig format
get.counts converts hdf5, bam, bigWig, and bigBed input files into R objects that can then be used as input to the main
multiseq function. In order to use
get.counts, make sure you have
bigWigToWig in your path (the last two are UCSC tools).
If /data/tools/ucsctools/ is your path to the UCSC tools and /usr/local/bin/ is your path to
samtools then add the following lines to your ~/.bashrc file:
Visualize results in the UCSC Genome Browser
multiseq input and output in the UCSC Genome Browser. In order to use them, you need a folder that is accessible via the Internet. If /some/path is the path to a folder that is accessible via the Internet - the "mountpoint" - and https:some/address is the http address of the mountpoint, you need to set the following shell environmental variables by adding the following lines to your ~/.bashrc:
# the following lines specify the mountpoint and the http address associated with the mountpoint export MOUNTPOINT_PATH="/some/path" export MOUNTPOINT_HTTP_ADDRESS="https:some/address"
After adding new lines to your .bashrc for the first time, remember to either login again, or do
Remember that when you submit jobs to a compute cluster (e.g. using SGE's qsub), they run in "batch mode" and may not execute your ~/.bashrc. To ensure that your jobs have the correct environmental variables set, you should be able to pass a flag to your cluster submission command (e.g. the
-V flag to